戴密斯·哈萨比斯：AI 的未来、模拟现实、物理与电子游戏

It’s hard for us humans to make any kind of clean predictions about highly nonlinear dynamical systems.

我们人类很难对高度非线性动力系统做出任何清晰的预测。

00:00 But again, to your point, we might be very surprised what classical learning systems might be able to do about even fluid.

但同样，就你的观点而言，我们可能会非常惊讶经典学习系统甚至能够对流体进行处理。

Yes, exactly.

是的，完全正确。

00:12 I mean, fluid dynamics, Navier-Stokes equations, these are traditionally thought of as very, very difficult intractable problems to do on classical systems.

我的意思是，流体动力学、纳维-斯托克斯方程，这些传统上被认为是在经典系统上解决非常非常困难的棘手问题。

They take enormous amounts of compute, weather prediction systems.

它们需要大量的计算和天气预报系统。

These kinds of things all involve fluid dynamics calculations.

这类事情都涉及流体动力学计算。

00:27 But again, if you look at something like Veo, our video generation model, it can model liquids quite well, surprisingly well.

但同样，如果你看看像我们的视频生成模型 Veo 这样的东西，它可以很好地模拟液体，效果出奇地好。

And materials, specular lighting, I love the ones where there’s people who generate videos where there’s clear liquids going through hydraulic presses and then it’s being squeezed out.

还有材质、镜面照明，我喜欢那些制作视频的人，其中有透明液体通过液压机，然后被挤出。

I used to write physics engines and graphics engines in my early days in gaming, and I know it’s just so painstakingly hard to build programs that can do that.

我在游戏行业的早期曾编写过物理引擎和图形引擎，而且我知道构建可以做到这一点的程序非常困难。

And yet somehow these systems are reverse engineering from just watching YouTube videos.

然而不知何故，这些系统只是通过观看 YouTube 视频进行逆向工程。

So presumably what’s happening is it’s extracting some underlying structure around how these materials behave.

因此，大概正在发生的事情是围绕这些材料的行为提取一些底层结构。

So perhaps there is some kind of lower dimensional manifold that can be learned if we actually fully understood what’s going on under the hood.

因此，如果我们真正完全理解幕后发生的事情，也许可以学习某种低维流形。

That’s maybe true of most of reality.

现实中的大部分情况也许都是如此。

The following is a conversation with Demis Hassabis, his second time on the podcast.

以下是与 Demis Hassabis 的对话，这是他第二次参加播客。

01:22 He is the leader of Google DeepMind and is now a Nobel Prize winner.

他是谷歌DeepMind的领导者，现在是诺贝尔奖获得者。

Demis is one of the most brilliant and fascinating minds in the world today working on understanding and building intelligence and exploring the big mysteries of our universe.

戴米斯是当今世界上最聪明、最迷人的人之一，致力于理解和构建智能以及探索宇宙的巨大奥秘。

This was truly an honor and a pleasure for me.

这对我来说确实是一种荣幸和快乐。

01:51 This is the Lex Fridman Podcast.

这是莱克斯·弗里德曼播客。

To support it, please check out our sponsors in the description and consider subscribing to this channel.

为了支持它，请查看说明中的赞助商并考虑订阅此频道。

And now, dear friends, here’s Demis Hassabis.

现在，亲爱的朋友们，这是黛米斯·哈萨比斯。

02:06 In your Nobel Prize lecture, you propose what I think is a super interesting conjecture that “any pattern that can be generated or found in nature can be efficiently discovered and modeled by a classical learning algorithm.” What kind of patterns or systems might be included in that?

在您的诺贝尔奖演讲中，您提出了一个我认为非常有趣的猜想：“自然界中可以生成或发现的任何模式都可以通过经典学习算法有效地发现和建模。”其中可能包含什么样的模式或系统？

Biology, chemistry, physics, maybe cosmology, neuroscience?

生物学、化学、物理学，也许还有宇宙学、神经科学？

What are we talking about?

我们在说什么？

Sure.

当然。

02:32 Well, look, I felt that it’s sort of a tradition, I think, of Nobel Prize lectures that you’re supposed to be a little bit provocative and I wanted to follow that tradition.

嗯，听着，我觉得这是诺贝尔奖演讲的一种传统，你应该有点挑衅性，我想遵循这个传统。

What I was talking about there is if you take a step back and you look at all the work that we’ve done, especially with the Alpha X projects, so I’m thinking AlphaGo, of course, AlphaFold, what they really are is we are building models of very combinatorially, high dimensional spaces that if you try to brute force a solution, find the best move and go, or find the exact shape of a protein, and if you enumerated all the possibilities, there wouldn’t be enough time in the time of the universe.

我所说的是，如果你退后一步，看看我们所做的所有工作，特别是 Alpha X 项目，所以我在想 AlphaGo，当然，AlphaFold，它们真正的本质是我们正在构建非常组合的高维空间模型，如果你试图强力解决方案，找到最佳的移动方式，或者找到蛋白质的确切形状，如果你列举了所有可能性，那么时间就不够了。宇宙。

03:08 So you have to do something much smarter.

所以你必须做一些更聪明的事情。

And what we did in both cases was build models of those environments and that guided the search in a smart way and that makes it tractable.

在这两种情况下，我们所做的都是构建这些环境的模型，并以智能的方式指导搜索，使其易于处理。

So if you think about protein folding, which is obviously a natural system, why should that be possible?

因此，如果你考虑一下蛋白质折叠，这显然是一个自然系统，为什么这可能呢？

How does physics do that?

物理学是如何做到这一点的？

Proteins fold in milliseconds in our bodies, so somehow physics solves this problem that we’ve now also solved computationally.

蛋白质在我们体内的折叠时间为几毫秒，因此物理学以某种方式解决了这个问题，我们现在也通过计算解决了这个问题。

And I think the reason that’s possible is that in nature, natural systems have structure because they were subject to evolutionary processes that shape them.

我认为这可能的原因是，在自然界中，自然系统具有结构，因为它们受到塑造它们的进化过程的影响。

And if that’s true, then you can maybe learn what that structure is.

如果这是真的，那么你也许可以了解这个结构是什么。

This perspective I think is a really interesting one.

我认为这个观点非常有趣。

03:49 You’ve hinted it at it, which is almost like crudely stated, anything that can be evolved can be efficiently modeled.

你已经暗示过，这几乎就像粗略地说，任何可以进化的东西都可以有效地建模。

Think there’s some truth to that?

你认为这有一定道理吗？

Yeah.

是的。

04:03 I sometimes call it survival of the stablest or something like that because of course there’s evolution for life, living things, but there’s also, if you think about geological times, so the shape of mountains, that’s been shaped by weathering processes over thousands of years, but then you can even take it cosmological, the orbits of planets, the shapes of asteroids.

我有时称之为“最稳定的生存”或类似的东西，因为当然有生命、生物的进化，但如果你考虑一下地质时代，那么还有山脉的形状，它是由数千年的风化过程塑造的，但你甚至可以把它理解为宇宙学，行星的轨道，小行星的形状。

These have all been survived kind of processes that have acted on them many, many times.

这些都经历了很多很多次的过程而幸存下来。

04:31 If that’s true, then there should be some sort of pattern that you can kind of reverse learn and a kind of manifold really that helps you search to the right solution, to the right shape and actually allow you to predict things about it in an efficient way because it’s not a random pattern.

如果这是真的，那么应该有某种你可以逆向学习的模式，以及一种真正可以帮助你搜索正确的解决方案、正确的形状的流形，并且实际上允许你以有效的方式预测有关它的事情，因为它不是随机模式。

So it may not be possible for man-made things or abstract things like factorizing large numbers because unless there’s patterns in the number space, which there might be, but if there’s not and it’s uniform, then there’s no pattern to learn, there’s no model to learn that will help you search.

因此，人造事物或抽象事物（例如分解大数）可能是不可能的，因为除非数字空间中存在模式（可能存在），但如果不存在并且它是统一的，那么就没有可学习的模式，也没有可以帮助您搜索的模型。

So you have to do brute force.

所以你必须用蛮力。

So in that case you maybe need a quantum computer, something like this.

所以在这种情况下你可能需要一台量子计算机，像这样的东西。

But in most things in nature that we’re interested in are not like that.

但我们感兴趣的自然界大多数事物并不是这样的。

They have structure that evolved for a reason and survived over time.

它们的结构是出于某种原因而进化并随着时间的推移而生存下来的。

And if that’s true, I think that’s potentially learnable by a neural network.

如果这是真的，我认为神经网络可能可以学习这一点。

It’s like nature is doing a search process and it’s so fascinating that in that search process, it’s creating systems that could be efficiently modeled.

就像大自然正在进行搜索过程一样，它是如此令人着迷，在这个搜索过程中，它正在创建可以有效建模的系统。

That’s right.

这是正确的。

05:31 Yeah.

是的。

So interesting.

太有趣了。

So they can be efficiently rediscovered or recovered because nature’s not random.

因此，它们可以被有效地重新发现或恢复，因为自然不是随机的。

05:33 Everything that we see around us, including the elements that are more stable, all of those things, they’re subject to some kind of selection process pressure.

我们在周围看到的一切，包括更稳定的元素，所有这些东西，它们都受到某种选择过程的压力。

Do you think because you’re also a fan of theoretical computer science and complexity, do you think we can come up with a complexity class, like a complexity zoo type of class where maybe it’s the set of learnable systems, the set of learnable natural systems, LNS.

你是否认为因为你也是理论计算机科学和复杂性的粉丝，你认为我们可以提出一个复杂性类，就像复杂性动物园类型的类，其中可能是一组可学习的系统，一组可学习的自然系统，LNS。

05:47 This is a Demis Hassabis new class of systems that could be actually learnable by classical systems in this kind of way, natural systems that can be modeled efficiently.

这是 Demis Hassabis 的一类新系统，实际上可以通过经典系统以这种方式学习，自然系统可以有效建模。

Yeah, I mean I’ve always been fascinated by the P equals NP question and what is model-able by classical systems, i.e. non-quantum systems, Turing machines in effect.

是的，我的意思是我一直对 P 等于 NP 问题以及经典系统（即非量子系统、图灵机）可建模的内容着迷。

06:17 And that’s exactly what I’m working on actually in my few moments of spare time with a few colleagues about should there be maybe a new class or problem that is solvable by this type of neural network process and kind of mapped onto these natural systems, so the things that exist in physics and have structure.

这正是我在业余时间和几位同事一起研究的问题，是否应该有一个新的类或问题可以通过这种类型的神经网络过程来解决，并映射到这些自然系统上，所以物理中存在并具有结构的东西。

So I think that could be a very interesting new way of thinking about it.

所以我认为这可能是一种非常有趣的新思考方式。

And it sort of fits with the way I think about physics in general, which is that I think information is primary, information is the most sort of fundamental unit of the universe, more fundamental than energy and matter.

这有点符合我对物理学的一般看法，即我认为信息是首要的，信息是宇宙最基本的单位，比能量和物质更基本。

I think they can all be converted into each other, but I think of the universe as a kind of informational system.

我认为它们都可以相互转换，但我认为宇宙是一种信息系统。

So when you think of the universe as an informational system, then the P equals NP question is a physics question.

因此，当你将宇宙视为一个信息系统时，那么 P 等于 NP 问题就是一个物理问题。

That’s right.

这是正确的。

And is a question that can help us actually solve the entirety of this whole thing going on.

这是一个可以帮助我们真正解决整个事件的问题。

Yeah, I think it’s one of the most fundamental questions actually if you think of physics as informational and the answer to that, I think it’s going to be very enlightening.

是的，我认为这实际上是最基本的问题之一，如果你将物理学视为信息性的，那么它的答案将非常具有启发性。

More specific to the P and MP question, again, some of the stuff we’re saying is kind of crazy right now just like the Christian Anfinsen Nobel Prize speech, controversial thing that he said sounded crazy and then you went and got a Nobel Prize for this with John Jumper, solved the problem.

更具体地说，关于 P 和 MP 问题，我们现在所说的一些东西有点疯狂，就像克里斯蒂安·安芬森诺贝尔奖演讲一样，他所说的有争议的事情听起来很疯狂，然后你和约翰·詹珀一起获得了诺贝尔奖，解决了问题。

07:29 So let me just stick to the P equals NP.

所以让我坚持 P 等于 NP。

Do you think there’s something in this thing we’re talking about that could be shown if you can do something like a polynomial time or constant time compute ahead of time and construct this gigantic model, then you can solve some of these extremely difficult problems in a theoretical computer science kind of way?

你认为我们正在谈论的这件事中是否有一些东西可以证明，如果你可以提前进行多项式时间或常数时间计算并构建这个巨大的模型，那么你可以用理论计算机科学的方式解决其中一些极其困难的问题？

Yeah, I think that there are actually a huge class of problems that could be couched in this way, the way we did AlphaGo and the way we did AlphaFold, where you model what the dynamics of the system is, the properties of that system, the environment that you are trying to understand, and then that makes the search for the solution or the prediction of the next step efficient.

是的，我认为实际上有一大类问题可以用这种方式来表达，就像我们做 AlphaGo 的方式和我们做 AlphaFold 的方式，你可以对系统的动力学是什么、系统的属性、你试图理解的环境进行建模，然后这使得对解决方案的搜索或下一步的预测变得高效。

08:12 Basically polynomial times, so tractable by a classical system, which a neural network is.

基本上是多项式时间，很容易被经典系统（神经网络）处理。

It runs on normal computers, right?

它可以在普通计算机上运行，​​对吗？

Classical computers, Turing machines in effect.

经典计算机，实际上是图灵机。

And I think it’s one of the most interesting questions there is, is how far can that paradigm go?

我认为这是最有趣的问题之一，这种范式能走多远？

08:53 I think we’ve proven, and the AI community in general that classical systems, Turing machines can go a lot further than we previously thought.

我认为我们和整个人工智能社区已经证明，经典系统、图灵机可以比我们之前想象的走得更远。

They can do things like model the structures of proteins and play go to better than world champion level.

他们可以做诸如蛋白质结构建模之类的事情，并且可以达到比世界冠军水平更好的水平。

And a lot of people would’ve thought maybe 10, 20 years ago that was decades away, or maybe you would need some sort of quantum machines to quantum systems to be able to do things like protein folding.

很多人可能会认为 10、20 年前那是几十年后的事了，或者也许你需要某种量子机器到量子系统才能完成蛋白质折叠等事情。

And so I think we haven’t really even sort of scratched the surface yet of what classical systems so-called could do.

所以我认为我们还没有真正触及所谓的经典系统所能做的事情的表面。

09:28 And of course, AGI being built on a neural network system on top of a neural network system on top of a classical computer would be the ultimate expression of that.

当然，建立在经典计算机之上的神经网络系统之上的通用人工智能将是这一点的最终体现。

And I think the limit, what the bounds of that kind of system, what it can do, it’s a very interesting question and directly speaks to the P equals NP question.

我认为极限，这种系统的界限是什么，它能做什么，这是一个非常有趣的问题，直接涉及到 P 等于 NP 问题。

What do you think, again, hypothetical, might be outside of this?

再次假设，您认为除此之外可能还有什么？

09:47 Maybe emergent phenomena?

也许是涌现现象？

If you look at cellular automata, you have extremely simple systems and then some complexity emerges.

如果你观察元胞自动机，你会发现系统非常简单，但随后就会出现一些复杂性。

Maybe that would be outside or even would you guess even that might be amenable to efficient modeling by a classical machine?

也许那是外部的，或者甚至你会猜测这可能适合经典机器的有效建模吗？

Yeah, I think those systems would be right on the boundary.

是的，我认为这些系统就在边界上。

10:09 So I think most emergent systems, cellular automata, things like that could be model-able by a classical system.

所以我认为大多数新兴系统、元胞自动机之类的东西都可以通过经典系统进行建模。

You just sort of do a forward simulation of it and it’d probably be efficient enough.

你只需对其进行前向模拟，它可能就足够有效了。

Of course there’s the question of things like chaotic systems where the initial conditions really matter and then you get to some uncorrelated end state.

当然，存在像混沌系统这样的问题，其中初始条件确实很重要，然后你会达到一些不相关的最终状态。

Now those could be difficult to model.

现在这些可能很难建模。

So I think these are kind of the open questions, but I think when you step back and look at what we’ve done with the systems and the problems that we’ve solved, and then you look at things like Veo 3 on video generation sort of rendering physics and lighting and things like that, really core fundamental things in physics, it’s pretty interesting.

所以我认为这些都是悬而未决的问题，但我认为当你退一步看看我们对系统做了什么以及我们已经解决的问题，然后你看看像 Veo 3 这样的视频生成类渲染物理和照明之类的东西，真正的物理学核心基础东西，这是非常有趣的。

I think it’s telling us something quite fundamental about how the universe is structured in my opinion.

我认为它告诉我们一些关于我认为宇宙是如何构造的非常基本的东西。

So in a way that’s what I want to build AGI for is to help us as scientists answer these questions like P equals NP.

所以在某种程度上，我想要建立 AGI 的目的是帮助我们科学家回答这些问题，比如 P 等于 NP。

Yeah, I think we might be continuously surprised about what is model-able by classical computers.

是的，我认为我们可能会对经典计算机可建模的内容不断感到惊讶。

11:09 I mean AlphaFold 3 on the interaction side is surprising that you can make any kind of progress on that direction.

我的意思是 AlphaFold 3 在交互方面令人惊讶，你可以在这个方向上取得任何进展。

AlphaGenome is surprising that you can map the genetic code to the function.

AlphaGenome 令人惊讶的是您可以将遗传密码映射到功能。

Kind of playing with the emergent kind of phenomena, you think there’s so many combinatorial options and then here you go, you can find the kernel that is efficiently model-able.

在处理新兴现象时，您认为有很多组合选项，然后您就可以找到可有效建模的内核。

Yes, because there’s some structure, there’s some landscape in the energy landscape or whatever it is that you can follow, some gradient you can follow.

是的，因为有一些结构，有一些能量景观中的景观或者任何你可以遵循的东西，一些你可以遵循的梯度。

11:36 And of course what neural networks are very good at is following gradients.

当然，神经网络非常擅长的是遵循梯度。

And so if there’s one to follow and you can specify the objective function correctly, you don’t have to deal with all that complexity, which I think is how we maybe have naively thought about it for decades, those problems.

因此，如果有一个可遵循的目标函数，并且您可以正确指定目标函数，那么您就不必处理所有这些复杂性，我认为这就是我们几十年来天真地思考这些问题的方式。

If you just enumerate all the possibilities, it looks totally intractable and there’s many, many problems like that.

如果你只是列举所有的可能性，看起来非常棘手，而且这样的问题还有很多很多。

12:06 And then you think, “Well, it’s like 10 to 300 possible protein structures, 10 to the 170 possible go positions.

然后你会想，“嗯，这就像 10 到 300 种可能的蛋白质结构，10 到 170 个可能的位置。

All of these are way more than atoms in the universe, so how could one possibly find the right solution or predict the next step?” But it turns out that it is possible.

所有这些都比宇宙中的原子要多得多，那么我们怎么可能找到正确的解决方案或预测下一步呢？”但事实证明这是可能的。

And of course reality in nature does do it.

当然，自然界的现实确实做到了这一点。

Proteins do fold.

蛋白质确实会折叠。

So that gives you confidence that there must be, if we understood how physics was doing that in a sense and we could mimic that process, i.e. model that process, it should be possible on our classical systems is basically what the conjecture is about.

因此，这让你相信，如果我们理解物理学在某种意义上是如何做到这一点的，并且我们可以模仿该过程，即对该过程进行建模，那么在我们的经典系统上应该是可能的，这基本上就是猜想的内容。

And of course there’s nonlinear dynamical systems, highly nonlinear dynamical systems, everything involving fluid.

当然，还有非线性动力系统、高度非线性动力系统，一切都涉及流体。

12:44 I recently had a conversation with Terence Tao who mathematically contends with a very difficult aspect of systems that have some singularities in them that break the mathematics, and it’s just hard for us humans to make any kind of clean predictions about highly nonlinear dynamical systems.

我最近与陶哲轩进行了一次对话，他在数学上与系统的一个非常困难的方面进行了斗争，这些系统中存在一些打破数学的奇点，而我们人类很难对高度非线性动力系统做出任何清晰的预测。

But again, to your point, we might be very surprised what classical learning systems might be able to do about even fluid.

但同样，就你的观点而言，我们可能会非常惊讶经典学习系统甚至能够对流体进行处理。

Yes, exactly.

是的，完全正确。

13:16 I mean fluid dynamics, Navier-Stokes equations, these are traditionally thought of as very, very difficult, intractable kind of problems to do on classical systems.

我指的是流体动力学、纳维-斯托克斯方程，这些传统上被认为是在经典系统上解决非常非常困难的问题。

They take enormous amounts of compute, weather prediction systems.

它们需要大量的计算和天气预报系统。

These kind of things all involve fluid dynamics calculations.

这类事情都涉及到流体动力学计算。

But again, if you look at something like Veo, our video generation model, it can model liquids quite well, surprisingly well.

但同样，如果你看看像我们的视频生成模型 Veo 这样的东西，它可以很好地模拟液体，效果出奇地好。

And materials, specular lighting, I love the ones where there’s people who generate videos where there’s clear liquids going through hydraulic presses and then it’s being squeezed out.

还有材质、镜面照明，我喜欢那些制作视频的人，其中有透明液体通过液压机，然后被挤出。

I used to write physics engines and graphics engines in my early days in gaming, and I know it’s just so painstakingly hard to build programs that can do that.

我在游戏行业的早期曾编写过物理引擎和图形引擎，而且我知道构建可以做到这一点的程序非常困难。

And yet somehow these systems are reverse engineering from just watching YouTube videos.

然而不知何故，这些系统只是通过观看 YouTube 视频进行逆向工程。

So presumably what’s happening is it’s extracting some underlying structure around how these materials behave.

因此，大概正在发生的事情是围绕这些材料的行为提取一些底层结构。

So perhaps there is some kind of lower dimensional manifold that can be learned if we actually fully understood what’s going on under the hood.

因此，如果我们真正完全理解幕后发生的事情，也许可以学习某种低维流形。

That’s maybe true of most of reality.

现实中的大部分情况也许都是如此。

Yeah, I’ve been continuously precisely by this aspect of Veo 3.

是的，我一直在关注 Veo 3 的这一方面。

14:26 I think a lot of people highlight different aspects including the comedic and the mean and all that kind of stuff.

我认为很多人都强调不同的方面，包括喜剧和卑鄙等等。

And then the ultra realistic ability to capture humans in a really nice way that’s compelling and feels close to reality, and then combine that with native audio.

然后是超现实的能力，以一种非常好的方式捕捉人类，这种方式令人信服并且感觉接近现实，然后将其与原生音频相结合。

All of those are marvelous things about Veo 3, but exactly the thing you’re mentioning, which is the physics.

所有这些都是 Veo 3 的奇妙之处，但正是你所提到的，那就是物理学。

Yeah.

是的。

It’s not perfect, but it’s pretty damn good.

它并不完美，但已经非常好了。

14:53 And then the really interesting scientific question is what is it understanding about our world in order to be able to do that?

然后，真正有趣的科学问题是，为了能够做到这一点，它对我们的世界有什么了解？

Because if the cynical, take with diffusion models, there’s no way to understands anything.

因为如果愤世嫉俗地采用扩散模型，就无法理解任何事情。

But I don’t think you can generate that kind of video without understanding.

但我认为你无法在不理解的情况下生成那种视频。

And then our own philosophical notion of what it means to understand then is brought to the surface.

然后我们自己关于理解那时意味着什么的哲学概念就浮出水面。

To what degree do you think Veo 3 understands our world?

您认为 Veo 3 在多大程度上了解我们的世界？

I think to the extent that it can predict the next frames in a coherent way, that is a form of understanding, not in the anthropomorphic version of, it’s not some kind of deep philosophical understanding of what’s going on, I don’t think these systems have that, but they certainly have modeled enough of the dynamics, put it that way, that they can pretty accurately generate whatever it is, eight seconds of consistent video that by eye, at least at a glance, is quite hard to distinguish what the issues are.

我认为它可以以连贯的方式预测下一帧，这是一种理解形式，不是拟人化的版本，不是对正在发生的事情的某种深刻的哲学理解，我不认为这些系统有这种能力，但他们肯定已经对动态进行了足够的建模，这样说吧，他们可以非常准确地生成无论是什么，八秒的一致视频，通过眼睛，至少乍一看，很难区分问题是什么。

15:57 And imagine that in two or three more years’ time, that’s the thing I’m thinking about and how incredible that will look given where we’ve come from the early versions of that one or two years ago.

想象一下，再过两三年，这就是我正在考虑的事情，考虑到我们在一两年前的早期版本，这将看起来多么令人难以置信。

And so the rate of progress is incredible.

所以进步的速度是令人难以置信的。

And I think I’m like you is like a lot of people love all of the stand-up comedians and that actually captures a lot of human dynamics very well and body language, but actually the thing I’m most impressed with and fascinated by is the physics behavior, the lighting and materials and liquids.

我想我就像你们一样，很多人都喜欢所有的单口喜剧演员，它们实际上很好地捕捉了很多人类动态和肢体语言，但实际上，我印象最深刻和最着迷的是物理行为、灯光、材料和液体。

And it’s pretty amazing that it can do that.

它能做到这一点真是太神奇了。

And I think that shows that it has some notion of at least intuitive physics, how things are supposed to work intuitively, maybe the way that a human child would understand physics, right, as opposed to a PhD student really being able to unpack all the equations.

我认为这表明它至少有一些直观的物理概念，事物应该如何直观地工作，也许是人类孩子理解物理的方式，对吧，而不是博士生真正能够解开所有方程。

It’s more of an intuitive physics understanding.

这更多的是一种直观的物理理解。

Well, that intuitive physics understanding, that’s the base layer, that’s the thing people sometimes call a common-sense.

嗯，直观的物理理解，这是基础层，这就是人们有时所说的常识。

16:53 It really understands something.

确实是明白了一些东西。

I think that really surprised a lot of people.

我认为这确实让很多人感到惊讶。

It blows my mind that I just didn’t think it would be possible to generate that level of realism without understanding.

令我震惊的是，我只是认为在不理解的情况下不可能产生那种程度的现实主义。

There’s this notion that you can only understand the physical world by having an embodied AI system, a robot that interacts with that world.

有一种观点认为，只有拥有具体的人工智能系统（即与物理世界交互的机器人）才能理解物理世界。

That’s the only way to construct an understanding of that world.

这是构建对这个世界的理解的唯一方法。

But Veo 3 is directly challenging that it feels like.

但 Veo 3 给人的感觉是直接挑战。

Yes, and it’s very interesting, even if you were to ask me five, 10 years ago, I would’ve said, even though I was immersed in all of this, I would’ve said, “Well, yeah, you probably need to understand intuitive physics.

是的，这非常有趣，即使你在五、十年前问我，我也会说，即使我沉浸在这一切之中，我也会说，“嗯，是的，你可能需要理解直觉物理学。

17:27 If I push this off the table, this glass, it will maybe shatter and the liquid will spill out.

如果我把它从桌子上推开，这个玻璃杯，它可能会碎裂，液体会溢出。

So we know all of these things.” But I thought that, and there’s a lot of theories in neuroscience, it’s called action in perception where you need to act in the world to really, truly perceive it in a deep way.

所以我们知道所有这些事情。”但我认为，神经科学中有很多理论，这被称为感知中的行动，你需要在世界上采取行动，才能真正、真正地深刻地感知它。

And there was a lot of theories about you’d need embodied intelligence or robotics or something, or maybe at least simulated action so that you would understand things like intuitive physics.

有很多理论说你需要具身智能或机器人技术之类的东西，或者至少需要模拟动作，这样你才能理解直觉物理学等东西。

18:06 But it seems like you can understand it through passive observation, which is pretty surprising to me.

但似乎你可以通过被动观察来理解它，这让我很惊讶。

And again, I think hints at something underlying about the nature of reality in my opinion, beyond just the cool videos that it generates.

再说一次，我认为在我看来，这暗示了现实本质的一些潜在的东西，而不仅仅是它生成的很酷的视频。

And of course there’s next stages is maybe even making those videos interactive so one can actually step into them and move around them, which would be really mind-blowing, especially given my games background.

当然，下一阶段甚至可能使这些视频具有互动性，这样人们就可以真正走进它们并在它们周围移动，这将非常令人兴奋，特别是考虑到我的游戏背景。

So you can imagine.

所以你可以想象。

And then I think we’re starting to get towards what I would call a world model, a model of how the world works, the mechanics of the world, the physics of the world, and the things in that world.

然后我认为我们开始走向我所说的世界模型，世界如何运作、世界的力学、世界的物理以及那个世界中的事物的模型。

And of course that’s what you would need for a true AGI system.

当然，这就是真正的 AGI 系统所需要的。

I have to talk to you about video games.

我必须和你谈谈电子游戏。

18:50 So you are being a bit trolley.

所以你有点手推车。

I think you’re having more and more fun on Twitter, on X, which is great to see.

我认为你在 Twitter、X 上玩得越来越开心，这很高兴看到。

So a guy named Jimmy Apples tweeted, “Let me play a video game of my Veo 3 videos already.

因此，一个名叫 Jimmy Apples 的人在推特上写道：“让我玩一下我的 Veo 3 视频的视频游戏吧。

Google cooked so good.

谷歌煮得真好。

Playable world models wen?” And then you co-tweeted that with, “Now, wouldn’t that be something?” So how hard is it to build game worlds with AI?

可玩的世界模型是吗？”然后你在推特上写道：“现在，这不是吗？”那么用人工智能构建游戏世界有多难呢？

Maybe can you look out into the future feature of video games five, 10 years out?

也许您能展望一下五年、十年后电子游戏的未来特征吗？

What do you think that looks like?

你觉得那看起来像什么？

Well, games were my first love really.

嗯，游戏确实是我的初恋。

19:27 And doing AI for games was the first thing I did professionally in my teenage years and with the first major AI systems that I built and I always want to scratch that itch one day and come back to that.

为游戏开发人工智能是我在青少年时期专业做的第一件事，也是我构建的第一个主要人工智能系统，我总是希望有一天能解决这个问题，然后再回来。

And I will do, I think, and I think I sort of dream about what would I have done back in the nineties if I’d had access to the kind of AI systems we have today?

我想，我会做的，我想，我想，如果我能够使用我们今天拥有的那种人工智能系统，我会在九十年代做什么？

And I think you could build absolutely mind-blowing games.

我认为你可以构建绝对令人兴奋的游戏。

19:55 And I think the next stage is I always used to love making, all the games I’ve made are open world games, so they’re games where there’s a simulation and then there’s AI characters, and then the player interacts with that simulation and the simulation adapts to the way the player plays.

我认为下一阶段是我一直习惯于制作，我制作的所有游戏都是开放世界游戏，所以它们是有模拟的游戏，然后有人工智能角色，然后玩家与模拟互动，模拟会适应玩家的游戏方式。

And I always thought they were the coolest games because, so games like Theme Park that I worked on where everybody’s game experience would be unique to them because you’re kind of co-creating the game.

我一直认为它们是最酷的游戏，因为我制作的像主题公园这样的游戏，每个人的游戏体验都是独一无二的，因为你们是共同创造游戏的。

We set up the parameters, we set up initial conditions, and then you as the player immersed in it, and then you are co-creating it with the simulation.

我们设置参数，设置初始条件，然后你作为玩家沉浸其中，然后你与模拟共同创建它。

But of course it’s very hard to program open world games.

当然，开发开放世界游戏是非常困难的。

You’ve got to be able to create content whichever direction the player goes in, and you want it to be compelling no matter what the player chooses.

无论玩家进入哪个方向，您都必须能够创建内容，并且无论玩家选择什么，您都希望它具有吸引力。

And so it was always quite difficult to build things like cellular automata actually, type of those kind of classical systems, which created some emergent behavior, but they’re always a little bit fragile, a little bit limited.

因此，实际上构建像元胞自动机这样的东西总是相当困难的，这种经典系统的类型会产生一些紧急行为，但它们总是有点脆弱，有点有限。

Now we are maybe on the cusp in the next few years, five, 10 years of having AI systems that can truly create around your imagination, can dynamically change the story and storytell the narrative around and make it dramatic no matter what you end up choosing.

现在，我们可能正处于未来几年、五年、十年的风口浪尖，人工智能系统可以真正围绕你的想象力进行创造，可以动态地改变故事并讲述周围的叙述，并使其变得戏剧化，无论你最终选择什么。

So it’s like the ultimate choose your own adventure sort of game.

所以这就像终极选择你自己的冒险游戏。

And I think maybe we are within reach, if you think of a kind of interactive version of Veo and then wind that forward five to 10 years and imagine how good it’s going to be.

我认为，如果你想到一种 Veo 的交互式版本，然后将其向前推进 5 到 10 年，想象一下它会有多好，也许我们就可以实现。

Yeah.

是的。

21:24 So you said a lot of super interesting stuff there.

所以你在那里说了很多超级有趣的事情。

So one, the open world, built into that is a deep personalization the way you’ve described it.

因此，开放世界是一种深度个性化，就像你所描述的那样。

So it’s not just that it’s open world, that you can open any door and there’ll be something there, it’s that it’s the choice of which door you open in an unconstrained way defines the worlds you see.

所以，这不仅仅是因为它是一个开放的世界，你可以打开任何一扇门，那里就会有一些东西，而是你以不受限制的方式打开哪扇门的选择定义了你所看到的世界。

So some games try to do that, they give you choice, but it’s really just an illusion of choice because you only, like Stanley Parable, a game I actually played, it’s really, there’s a couple of doors and it really just takes you down a narrative.

所以有些游戏试图做到这一点，他们给你选择，但这实际上只是一种选择的幻觉，因为你只是，就像《斯坦利寓言》一样，我实际上玩过一款游戏，它真的有几扇门，它真的只是让你记下一个故事。

Stanley Parable is a great video game.

史丹利寓言是一款很棒的视频游戏。

I recommend people play it, that in a meta way, mocks the illusion of choice, and there’s philosophical notions of free will and so on.

我建议人们玩它，以元的方式嘲笑选择的幻觉，并且有自由意志的哲学概念等等。

22:12 But I do, one of my favorite games of Elder Scrolls is Daggerfall I believe, that they really played with a random generation of the dungeons of if you can step in and they give you this feeling of an open world.

但我确实，我最喜欢的《上古卷轴》游戏之一是《Daggerfall》，我相信他们真的玩了随机生成的地下城，如果你能介入，他们会给你一种开放世界的感觉。

And there you mentioned interactivity.

你在那里提到了交互性。

You don’t need to interact.

你不需要互动。

That’s the first step because you don’t need to interact that much.

这是第一步，因为您不需要进行太多互动。

You just, when you open the door, whatever you see is randomly generated for you.

你只是，当你打开门时，你看到的任何东西都是随机生成的。

And that’s already an incredible experience because you might be the only person to ever see that.

这已经是一次令人难以置信的体验，因为您可能是唯一见过这种情况的人。

Yeah, exactly.

是的，完全正确。

22:46 But what you’d like is a little bit better than just sort of a random generation.

但你想要的比随机生成要好一点。

So you’d like, and also better than a simple AB hard coded choice, right?

所以你会想要，而且比简单的 AB 硬编码选择更好，对吧？

That’s not really open world, as you say.

正如你所说，这并不是真正的开放世界。

It’s just giving you the illusion of choice.

它只是给你一种选择的错觉。

What you want to be able to do is potentially anything in that game environment.

您想要做的可能是该游戏环境中的任何事情。

And I think the only way you can do that is to have generated systems, systems that will generate that on the fly.

我认为你能做到这一点的唯一方法就是拥有生成系统，能够动态生成它的系统。

Of course, you can’t create infinite amounts of game assets.

当然，你不可能创造无限数量的游戏资产。

It’s expensive enough already how AAA games are made today.

如今制作 AAA 级游戏的成本已经够高的了。

And that was obvious to us back in the nineties when I was working on all these games.

早在九十年代，当我开发所有这些游戏时，这一点对我们来说就很明显了。

23:25 I think maybe Black & White was the game that I worked on early stages of that, that had still probably the best AI, learning AI, in it.

我想也许《黑与白》是我在早期阶段开发的游戏，它可能仍然拥有最好的人工智能，即学习人工智能。

It was an early reinforcement learning system that you were looking after this mythical creature and growing it and nurturing it.

这是一个早期的强化学习系统，你正在照顾这个神话生物，培育它并培育它。

And depending how you treated it, it would treat the villagers in that world the same way.

而根据你如何对待它，它也会以同样的方式对待那个世界的村民。

So if you were mean to it, it would be mean.

所以如果你对它很刻薄，那就很刻薄了。

If you were good, it would be protective.

如果你表现好的话，就会有保护作用。

And so it was really a reflection of the way you played it.

所以这实际上反映了你的演奏方式。

So actually all of the, I’ve been working on simulations and AI through the medium of games at the beginning of my career, and really the whole of what I do today, it’s still a follow on from those early more hard coded ways of doing the AI to now fully general learning systems that are trying to achieve the same thing.

所以实际上，在我职业生涯的初期，我一直在通过游戏为媒介来研究模拟和人工智能，而我今天所做的全部工作，仍然是从那些早期的更硬编码的人工智能方法到现在试图实现同样目标的完全通用的学习系统的延续。

Yeah, it is been interesting, hilarious, and fun to watch you and Elon obviously itching to create games because you’re both gamers.

是的，看到你和埃隆显然渴望创造游戏，这真是有趣、搞笑、有趣，因为你们都是游戏玩家。

24:12 And one of the sad aspects of your incredible success in so many domains of science, like serious adult stuff, that you might not have time to really create a game, you might end up creating the tooling that others will create the game.

你在如此多的科学领域（比如严肃的成人事物）取得了令人难以置信的成功，其中一个可悲的方面是，你可能没有时间真正创建游戏，你最终可能会创建其他人将创建游戏的工具。

You have to watch others create the thing you’ve always dreamed of.

你必须看着别人创造你一直梦想的东西。

Do you think it’s possible you can somehow in your extremely busy schedule actually find time to create something like Black & White, an actual video game where you could make the childhood dream become reality?

你是否认为你有可能在极其繁忙的日程中找到时间来创造像《黑与白》这样的东西，一个真正的视频游戏，让你童年的梦想成为现实？

Well, there’s two things where I think about that is maybe with vibe coding as it gets better and there’s a possibility that I could, one could do that actually in your spare time.

嗯，我想到了两件事，也许随着振动编码变得更好，我有可能可以在业余时间真正做到这一点。

24:58 So I’m quite excited about that as that would be my project if I got the time to do some vibe coding.

所以我对此感到非常兴奋，因为如果我有时间做一些氛围编码，这将是我的项目。

I’m actually itching to do that.

我其实很想这么做。

And then the other thing is maybe it’s a sabbatical after AGI has been safely stewarded into the world and delivered into the world.

另一件事是，在 AGI 被安全地管理并交付给世界之后，这可能是一个休假。

That, and then working on my physics theory as we talked about at the beginning, those would be my two post-AGI projects, let’s call it that way.

然后，正如我们在开始时讨论的那样，研究我的物理理论，这将是我的两个后 AGI 项目，让我们这样称呼它。

I would love to see which you choose, solving the problem that some of the smartest people in human history contended with, P equals NP, or creating a cool video game.

我很想看看你会选择哪一个，是解决人类历史上一些最聪明的人所面临的问题，P 等于 NP，还是创造一款很酷的视频游戏。

But in my world, they’d be related because it would be an open world simulated game as realistic as possible.

但在我的世界里，它们是相关的，因为这将是一个尽可能真实的开放世界模拟游戏。

25:44 So what is the universe?

那么宇宙是什么？

That’s speaking to the same question and P equals NP.

这是同一个问题，P 等于 NP。

I think all these things are related, at least in my mind.

我认为所有这些事情都是相关的，至少在我看来是这样。

I mean in a really serious way, video games sometimes are looked down upon as just this fun side activity.

我的意思是，以一种非常严肃的方式，电子游戏有时被视为只是一种有趣的副业活动。

25:59 But especially as AI does more and more of the difficult, boring tasks, something we in modern world called work, video games is the thing in which we may find meaning, in which we may find what to do with our time.

但尤其是当人工智能承担越来越多困难、无聊的任务时，我们在现代世界称之为工作，视频游戏是我们可以找到意义的东西，我们可以在其中找到如何打发时间的东西。

You could create incredibly rich, meaningful experiences.

您可以创造极其丰富、有意义的体验。

That’s what human life is.

这就是人类的生活。

And then in video games, you can create more sophisticated, more diverse ways of living.

然后在视频游戏中，你可以创造更复杂、更多样化的生活方式。

Right?

正确的？

That’s the point?

这就是重点吗？

I think so.

我想是的。

26:41 I mean, those of us who love games and I still do is it’s almost can let your imagination run wild, right?

我的意思是，对于我们这些热爱游戏的人来说，我仍然喜欢它，因为它几乎可以让你的想象力尽情发挥，对吗？

I used to love games and working on games so much because it’s the fusion, especially in the nineties and early two thousands, the sort of golden era and maybe the eighties of the games industry.

我曾经非常喜欢游戏并从事游戏工作，因为它是融合，特别是在九十年代和两千世纪初，游戏行业的黄金时代，也许是八十年代。

And it was all being discovered.

而这一切都被发现了。

New genres were being discovered.

新的流派不断被发现。

We weren’t just making games, we felt we were creating a new entertainment medium that never existed before.

我们不只是制作游戏，我们觉得我们正在创造一种以前从未存在过的新娱乐媒体。

Especially with these open world games and simulation games where you, as the player, were co-creating the story.

尤其是在这些开放世界游戏和模拟游戏中，你作为玩家，正在共同创造故事。

There’s no other media, entertainment media, where you do that, where you as the audience actually co-create the story.

没有其他媒体、娱乐媒体可以让你作为观众真正共同创造故事。

27:23 And of course now with multiplayer games as well, it can be a very social activity and can explore all kinds of interesting worlds in that.

当然，现在也有了多人游戏，它可以是一项非常社交的活动，可以探索其中各种有趣的世界。

But on the other hand, it’s very important to also enjoy and experience the physical world.

但另一方面，享受和体验物质世界也非常重要。

But the question is then I think we’re going to have to confront the question again of what is the fundamental nature of reality?

但问题是，我认为我们将不得不再次面对这样的问题：现实的基本本质是什么？

What is going to be the difference between these increasingly realistic simulations and multiplayer ones and emergent and what we do in the real world?

这些越来越现实的模拟和多人游戏以及新兴的模拟和我们在现实世界中所做的事情之间会有什么区别？

Yeah, there’s clearly a huge amount of value to experiencing the real world, nature.

是的，体验现实世界、自然显然具有巨大的价值。

27:55 There’s also a huge amount of value in experiencing other humans directly in person the way we’re sitting here today, but we need to really scientifically rigorously answer the question why and which aspect of that can be mapped into the virtual world.

以我们今天坐在这里的方式直接亲身体验其他人类也具有巨大的价值，但我们需要真正科学地、严格地回答为什么以及哪些方面可以映射到虚拟世界的问题。

Exactly.

确切地。

It’s not enough to say, “Yeah, you should go touch grass and hang out in nature.” It’s like why exactly is that valuable?

仅仅说“是的，你应该去触摸草地并在大自然中闲逛”是不够的。就像为什么那么有价值？

Yes.

是的。

28:25 And I guess that’s maybe the thing that’s been haunting me or obsessing me from the beginning of my career.

我想这可能是从我职业生涯之初就一直困扰着我或让我着迷的事情。

If you think about all the different things I’ve done, they’re all related in that way.

如果你想想我做过的所有不同的事情，它们都是以这种方式相关的。

The simulation, nature of reality, and what is the bounds of what can be modeled.

模拟、现实的本质以及可建模的范围。

Sorry for the ridiculous question, but so far, what is the greatest video game of all time?

抱歉问了这个荒谬的问题，但到目前为止，有史以来最伟大的视频游戏是什么？

28:40 What’s up there?

那里有什么？

Well, my favorite one of all time is Civilization, I have to say.

嗯，我不得不说，我最喜欢的就是《文明》。

28:45 That was the Civilization I and Civilization II, my favorite games of all time.

那就是《文明一》和《文明二》，我一直以来最喜欢的游戏。

I can only assume you’ve avoided the most recent one because it would probably, that would be your sabbatical.

我只能假设你已经避开了最近的一次，因为那很可能是你的休假。

28:54 You would disappear.

你就会消失。

Yes, exactly.

是的，完全正确。

29:03 They take a lot of time, these Civilization games, so I’ve got to be careful with them.

这些《文明》游戏需要花费很多时间，所以我必须小心对待它们。

Fun question.

有趣的问题。

29:08 You and Elon seem to be somehow solid gamers.

你和埃隆似乎都是可靠的游戏玩家。

Is there a connection between being great at gaming and being great leaders of AI companies?

擅长游戏和成为人工智能公司的伟大领导者之间有联系吗？

I don’t know.

我不知道。

29:20 It’s an interesting one.

这是一件有趣的事。

I mean, we both love games and it’s interesting, he wrote games as well to start off with.

我的意思是，我们都喜欢游戏，这很有趣，他一开始也写游戏。

Probably, especially in the era I grew up in where home computers just became a thing in the late eighties and nineties, especially in the UK, I had a spectrum and then a Commodore Amiga 500, which was my favorite computer ever.

可能，尤其是在我成长的时代，家用电脑在八十年代末和九十年代才刚刚流行，尤其是在英国，我有一个频谱，然后是 Commodore Amiga 500，这是我最喜欢的电脑。

And that’s why I learned all my programming.

这就是我学习所有编程的原因。

And of course, it’s a very fun thing to program, is to program games.

当然，编程是一件非常有趣的事情，就是编写游戏。

So I think it’s a great way to learn programming, probably still is.

所以我认为这是学习编程的好方法，可能仍然是。

And then of course, I immediately took it in directions of AI and simulations, so I was able to express my interest in games and my wider scientific interests all together.

当然，我立即将其转向人工智能和模拟方向，因此我能够同时表达我对游戏的兴趣和更广泛的科学兴趣。

30:00 And my sort of wider scientific interests all together.

还有我更广泛的科学兴趣。

And then the final thing I think that’s great about games is it fuses artistic design, art, with the most cutting edge programming.

我认为游戏的最后一点是它融合了艺术设计、艺术和最前沿的编程。

So again, in the nineties, all of the most interesting technical advances were happening in gaming, whether that was AI, graphics, physics engines, hardware, even GPUs of course were designed for gaming originally.

再说一次，在九十年代，所有最有趣的技术进步都发生在游戏领域，无论是人工智能、图形、物理引擎、硬件，当然甚至 GPU 最初都是为游戏设计的。

So everything that was pushing computing forward in the nineties was due to gaming.

因此，九十年代推动计算发展的一切都归功于游戏。

So interestingly, that was where the forefront of research was going on and it was this incredible fusion with art.

有趣的是，这就是最前沿的研究正在进行的地方，也是与艺术令人难以置信的融合。

Graphics, but also music, and just the whole new media of storytelling.

图形，还有音乐，以及讲故事的全新媒体。

And I love that.

我喜欢这样。

For me, it’s this sort of multidisciplinary kind of effort is again something I’ve enjoyed my whole life.

对我来说，这种多学科的努力又是我一生所享受的。

I have to ask you, I almost forgot about one of the many, and I would say one of the most incredible things recently that somehow didn’t yet get enough attention is AlphaEvolve.

我不得不问你，我几乎忘记了其中之一，我想说最近最令人难以置信的事情之一，但不知何故尚未得到足够的关注，那就是 AlphaEvolve。

30:52 We talked about Evolution a little bit, but it’s the Google DeepMind system that evolves algorithms.

我们稍微讨论了进化，但进化算法的是 Google DeepMind 系统。

Are these kinds of Evolution-like techniques promising as a component of future super intelligence systems?

这些类似进化的技术是否有希望成为未来超级智能系统的组成部分？

So for people who don’t know, it’s kind of, I don’t know if it’s fair to say it’s LLM guided Evolution search because Evolution algorithms are doing the search and LLMs are telling you where.

所以对于那些不知道的人来说，我不知道说这是法学硕士引导的进化搜索是否公平，因为进化算法正在进行搜索，法学硕士告诉你在哪里。

Yes.

是的。

31:28 Yes, exactly.

是的，完全正确。

So LLMs are kind of proposing some possible solutions and then you use evolutionary computing on top to find some novel part of the search space.

所以法学硕士是提出一些可能的解决方案，然后你在上面使用进化计算来找到搜索空间的一些新颖部分。

So actually I think it’s an example of very promising directions where you combine LLMs or foundation models with other computational techniques.

所以实际上我认为这是一个非常有前途的方向的例子，你可以将法学硕士或基础模型与其他计算技术相结合。

Evolutionary methods is one, but you could also imagine Monte Carlo tree search.

进化方法是其中之一，但您也可以想象蒙特卡罗树搜索。

Basically many types of search algorithms or reasoning algorithms sort of on top of or using the foundation models as a basis.

基本上，许多类型的搜索算法或推理算法都建立在基础模型之上或使用基础模型作为基础。

So I actually think there’s quite a lot of interesting things to be discovered probably with these sort of hybrid systems, let’s call them.

所以我实际上认为，通过这些混合系统，我们可能会发现很多有趣的事情，让我们称之为混合系统。

But not to romanticize Evolution.

但不要将进化论浪漫化。

Yeah.

是的。

I’m only human, but you think there’s some value in whatever that mechanism is?

我只是人类，但你认为无论这个机制是什么都有其价值吗？

32:12 Because we already talked about natural systems.

因为我们已经讨论过自然系统。

Do you think where there’s a lot of low-hanging fruit of us understanding, being able to model, being able to simulate Evolution and then using that, whatever we understand about that nature-inspired mechanism, to then do search better and better and better?

你认为我们理解、能够建模、能够模拟进化然后利用它，无论我们对自然启发机制的理解是什么，有很多唾手可得的成果，然后越来越好地进行搜索？

Yes.

是的。

32:36 So if you think about, again, breaking down the solar systems we’ve built to their really fundamental core, you’ve got the model of the underlying dynamics of the system.

因此，如果你再次考虑一下，将我们建造的太阳系分解到其真正的基本核心，你就会得到系统底层动力学的模型。

And then if you want to discover something new, something novel that hasn’t been seen before, then you need some kind of search process on top to take you to a novel of the search space.

然后，如果你想发现一些新的东西，一些以前从未见过的新颖的东西，那么你需要某种顶部的搜索过程来带你进入搜索空间的小说。

And you can do that in a number of ways.

您可以通过多种方式做到这一点。

Evolutionary computing is one.

进化计算就是其中之一。

With AlphaGo, we just use Monte Carlo Tree Search and that’s what found move 37, the new never seen before strategy in Go.

对于 AlphaGo，我们只使用蒙特卡罗树搜索，这就是找到第 37 步的方法，这是围棋中以前从未见过的新策略。

And so that’s how you can go beyond potentially what is already known.

这就是你如何能够超越已知的范围。

So the model can model everything that you currently know about, all the data that you currently have.

因此，该模型可以对您当前了解的所有内容以及您当前拥有的所有数据进行建模。

But then how do you go beyond that?

但你如何超越这个范围呢？

So that starts to speak about the ideas of creativity.

那么这就开始谈论创造力的想法了。

33:28 How can these systems create something new?

这些系统如何创造新事物？

In fact discover something new?

事实上发现一些新东西吗？

Obviously this is super relevant for scientific discovery or pushing met science and medicine forward, which we want to do with these systems.

显然，这与科学发现或推动气象科学和医学向前发展非常相关，我们希望利用这些系统来实现这一点。

And you can actually bolt on some fairly simple search systems on top of these models and get you into a new region of space.

实际上，您可以在这些模型之上安装一些相当简单的搜索系统，然后进入一个新的空间区域。

Of course, you also have to make sure that you’re not searching that space totally randomly.

当然，您还必须确保您不是完全随机地搜索该空间。

It would be too big.

它太大了。

So you have to have some objective function that you’re trying to optimize and hill climb towards and that guides that search.

因此，你必须有一些目标函数，你要尝试优化和爬山，并指导搜索。

But there’s some mechanism of Evolution that are interesting maybe in the space of programs.

但在程序领域，进化的某些机制可能很有趣。

34:00 But then the space of programs an extremely important space, because you can probably generalize to everything.

但程序空间是一个极其重要的空间，因为你可能可以推广到一切。

But for example, mutation.

但例如，突变。

So it’s not just Monte Carlo Tree Search where it’s like a search.

因此，蒙特卡罗树搜索不仅仅是搜索。

You could every once in a while-

你可以每隔一段时间-

Combine things.

结合事物。

Combine things?

结合事物？

Yeah.

是的。

Things, like the components of a thing.

事物，就像事物的组成部分。

Yes.

是的。

So then what Evolution is really good at is not just the natural selection, it’s combining things and building increasingly complex hierarchical systems.

那么进化真正擅长的不仅仅是自然选择，它还结合事物并构建日益复杂的层次系统。

34:26 So that component is super interesting, especially with AlphaEvolve and the space of programs.

所以这个组件非常有趣，尤其是 AlphaEvolve 和程序空间。

Yeah, exactly.

是的，完全正确。

34:43 So you can get a bit of an extra property out of evolutionary systems, which is some new emerging capability may come about, right?

所以你可以从进化系统中获得一些额外的属性，这是可能会出现的一些新的能力，对吗？

Of course like happened with life, interestingly with naive, traditional evolutionary computing methods without LLMs and the modern AI, the problem with them, they were very well studied in the nineties and early two thousands and some promising results, but the problem was they could never work out how to evolve new properties, new emerging properties.

当然，就像生活中发生的那样，有趣的是，没有法学硕士和现代人工智能的天真、传统的进化计算方法，它们的问题是，它们在九十年代和两千世纪初得到了很好的研究，并取得了一些有希望的结果，但问题是他们永远无法弄清楚如何进化新的特性，新的新兴特性。

You always had a sort of subset of the properties that you put into the system, but maybe if we combine them with these foundation models, perhaps we can overcome that limitation.

你总是有一些属性的子集放入系统中，但如果我们将它们与这些基础模型结合起来，也许我们可以克服这个限制。

35:21 Obviously naturally evolution clearly did.

显然，自然进化显然做到了。

It did evolve new capabilities.

它确实发展出了新的能力。

So bacteria to where we are now.

所以细菌到了我们现在的位置。

So clearly that it must be possible with evolutionary systems to generate new patterns, going back to the first thing we talked about and new capabilities and emerging properties, and maybe we’re on the cusp of discovering how to do that.

很明显，进化系统一定有可能产生新的模式，回到我们谈论的第一件事以及新的能力和新兴的属性，也许我们正处于发现如何做到这一点的风口浪尖。

Yeah, listen, AlphaEvolve is one of the coolest things I’ve ever seen.

是的，听着，AlphaEvolve 是我见过的最酷的东西之一。

35:44 I’ve on my desk at home, most of my time is spent on that computer is just programming.

我在家里的办公桌上，大部分时间都花在了那台电脑上，只是编程。

And next to the three screens is a skull of a Tiktaalik, which is one of the early organisms that crawled out of the water onto land.

三个屏幕旁边是提塔利克鱼的头骨，提塔利克鱼是最早从水中爬到陆地上的生物之一。

And I just kind of watch that little guy.

我只是看着那个小家伙。

It’s like whatever the competition mechanism of Evolution is, it’s quite incredible.

就好像进化论的竞争机制是什么，这真是太不可思议了。

Yes.

是的。

It’s truly, truly incredible.

这真的非常非常令人难以置信。

36:17 Now whether that’s exactly the thing we need to do to do our search, but never dismiss the power of nature, what it did here.

现在，这是否正是我们进行搜索所需要做的事情，但永远不要忽视自然的力量，它在这里所做的一切。

And it’s amazing, which is a relatively simple algorithm, right?

令人惊奇的是，这是一个相对简单的算法，对吗？

36:27 Effectively, and it can generate all of this immense complexity emerges obviously running over 4 billion years of time.

实际上，它可以产生所有这些巨大的复杂性，显然是在 40 亿年的时间里出现的。

But you can think about that as again, a search process that ran over the physics substrate of the universe for a long amount of computational time, but then it generated all this incredible rich diversity.

但你可以再次思考这一点，一个搜索过程在宇宙的物理基础上运行了很长的计算时间，但随后它产生了所有这些令人难以置信的丰富多样性。

So many questions I want to ask you.

很多问题我想问你。

36:54 So one, you do have a dream, one of the natural systems you want to try to model is a cell.

所以，你确实有一个梦想，你想要尝试建模的自然系统之一是细胞。

That’s a beautiful dream.

那是一个美丽的梦。

I could ask you about that.

我可以问你这个问题。

I also just for that purpose on the AI scientist front just broadly, so there’s a essay from Daniel Cocotaglio, Scott Alexander and others that online steps along the way to get to ASI and it has a lot of interesting ideas in it, one of which is including a superhuman coder and a superhuman AI researcher.

我也正是出于这个目的，广泛地关注人工智能科学家前沿，所以 Daniel Cocotaglio、Scott Alexander 和其他人在网上写了一篇文章，介绍了通往 ASI 的过程，其中有很多有趣的想法，其中包括一个超人的编码员和一个超人的人工智能研究员。

And in that there’s a term of research taste that’s really interesting.

其中有一个非常有趣的研究品味术语。

So in everything you’ve seen, do you think it’s possible for AI systems to have research taste to help you in the way that AI co-scientist does, to help steer human brilliant scientists and then potentially by itself to figure out what are the directions where you want to generate truly novel ideas?

那么，在你所看到的一切中，你认为人工智能系统有可能具有研究品味，以人工智能联合科学家的方式帮助你，帮助引导人类杰出的科学家，然后有可能自己找出你想要产生真正新颖的想法的方向是什么？

That seems to be a really important component of how to do great science?

这似乎是如何进行伟大科学研究的一个非常重要的组成部分？

Yeah, I think that’s going to be one of the hardest things to mimic or model is this idea of taste or judgment.

是的，我认为品味或判断的想法将是最难模仿或模仿的事情之一。

38:03 I think that’s what separates the great scientists from the good scientists.

我认为这就是伟大科学家和优秀科学家的区别。

All professional scientists are good technically, otherwise they wouldn’t have made it that far in academia and things like that.

所有专业科学家的技术都很好，否则他们不会在学术界取得那么大的成就。

But then do you have the taste to sniff out what the right direction is, what the right experiment is, what the right question is?

但你是否有品味去嗅出什么是正确的方向、什么是正确的实验、什么是正确的问题？

So picking the right question is the hardest part of science and making the right hypothesis.

因此，选择正确的问题并做出正确的假设是科学中最困难的部分。

And that’s what today’s systems definitely they can’t do.

而这正是当今的系统绝对无法做到的。

So I often say it’s harder to come up with a conjecture, a really good conjecture than it is to solve it.

所以我经常说，提出一个猜想，一个真正好的猜想比解决它更难。

So we may have systems soon that can solve pretty hard conjectures.

因此，我们可能很快就会拥有可以解决相当困难的猜想的系统。

A maths Olympiad problems, where Alpha Proof last year our system got silver medal in that really hard problems.

去年，我们的系统在数学奥林匹克竞赛中的 Alpha Proof 难题中获得了银牌。

Maybe eventually we’ll better solve a Millennium Prize kind of problem.

也许最终我们会更好地解决千禧年奖之类的问题。

But could a system have come up with a conjecture worthy of study that someone like Terence Tao would’ve gone?

但是，一个系统能否提出一个值得研究的猜想，认为像陶哲轩这样的人会离开呢？

“You know what, that’s a really deep question about the nature of maths or the nature of numbers or the nature of physics.” And that is far harder type of creativity.

“你知道吗，这是一个关于数学本质、数字本质或物理本质的非常深刻的问题。”这是一种更难的创造力。

And we don’t really know.

我们真的不知道。

Today’s systems clearly can’t do that.

今天的系统显然无法做到这一点。

And we’re not quite sure what that mechanism would be.

我们不太确定这个机制是什么。

This kind of leap of imagination like Einstein had when he came up with special relativity and then general relativity with the knowledge he had at the time.

这种想象力的飞跃，就像爱因斯坦用当时的知识提出狭义相对论，然后提出广义相对论一样。

For conjecture, you want to come up with a thing that’s interesting, it’s amenable to proof?

对于猜想，你想提出一个有趣的东西，它可以被证明吗？

Yes.

是的。

So it’s easy to come up with a thing that’s extremely difficult.

所以想出一件极其困难的事情很容易。

39:42 It’s easy to come up with a thing that’s extremely easy, but at that very edge-

想出一件极其简单的事情很容易，但就在那个边缘——

That sweet spot of basically advancing the science and splitting the hypothesis space into two, ideally.

理想情况下，基本上推进科学发展并将假设空间一分为二的最佳点。

39:48 Right?

正确的？

Whether if it’s true or not true, you’ve learned something really useful and that’s hard.

不管这是真的还是假的，你都学到了一些真正有用的东西，但这很难。

And making something that’s also falsifiable and within the technologies that you currently have available.

并在你当前可用的技术范围内做出一些可证伪的东西。

So it’s a very creative process, actually.

所以这实际上是一个非常有创意的过程。

A highly creative process that I think just a kind of naive search on top of a model won’t be enough for that.

这是一个高度创造性的过程，我认为仅仅在模型之上进行简单的搜索是不够的。

The idea of splitting the hypothesis space in two is super interesting.

将假设空间一分为二的想法非常有趣。

40:21 So I’ve heard you say that there’s basically no failure or failure is extremely valuable if you construct the questions right, if you construct the experiments right, if you design them right, that failure or success are both useful, so perhaps because it splits the hypothesis basically in two, it’s like a binary search?

所以我听你说，如果你正确构建问题，如果你正确构建实验，如果你设计正确，那么基本上没有失败，或者失败是非常有价值的，失败或成功都是有用的，所以也许因为它基本上将假设一分为二，就像二分搜索？

Yes, that’s right.

是的，没错。

40:44 So when you do real Blue Sky research, there’s no such thing as failure really.

所以当你进行真正的蓝天研究时，真的不存在失败这样的事情。

As long as you are picking experiments and hypotheses that meaningfully split the hypothesis space and you learn something.

只要你选择有意义地分割假设空间的实验和假设，你就能学到一些东西。

You can learn something kind of equally valuable from an experiment that doesn’t work.

你可以从无效的实验中学到同样有价值的东西。

That should tell you if you’ve designed the experiment well and your hypotheses are interesting, it should tell you a lot about where to go next.

这应该告诉你，如果你的实验设计得很好，你的假设很有趣，它应该告诉你很多关于下一步该做什么的信息。

And then you’re effectively doing a search process and using that information in very helpful ways.

然后您就可以有效地进行搜索过程并以非常有用的方式使用该信息。

So to go to your dream of modeling a cell, what are the big challenges that lay ahead for us to make that happen?

那么，为了实现细胞建模的梦想，我们面临哪些巨大挑战？

41:17 We should maybe highlight that in AlphaFold, I mean there’s just so many leaps.

我们也许应该强调一点，在 AlphaFold 中，我的意思是有太多的飞跃。

So AlphaFold solved, if it’s fair to say, protein folding.

公平地说，AlphaFold 解决了蛋白质折叠问题。

And there’s so many incredible things we could talk about there, including the open sourcing, everything you’ve released AlphaFold 3 is doing protein, RNA, DNA interactions, which is super complicated and fascinating.

我们可以在那里谈论很多令人难以置信的事情，包括开源，你们发布的 AlphaFold 3 的所有内容都在进行蛋白质、RNA、DNA 相互作用，这是超级复杂和令人着迷的。

It’s amenable to modeling.

它适合建模。

AlphaGenome predicts how small genetic changes if we think about single mutations, how they link to actual function.

如果我们考虑单个突变，AlphaGenome 可以预测基因变化有多小，以及它们如何与实际功能联系起来。

So it seems like it’s creeping along to sophisticated to much more complicated things like a cell.

因此，它似乎正在逐渐变得复杂，变得更加复杂，比如细胞。

But a cell has a lot of really complicated components.

但细胞有很多非常复杂的组件。

So what I’ve tried to do throughout my career is I have these really grand dreams and then I try to, as you’ve noticed, but I try to break them down.

因此，在我的整个职业生涯中，我一直在努力做的是，我有这些非常宏伟的梦想，然后我尝试着，正如你所注意到的，但我试图打破它们。

42:11 It’s easy to have a kind of crazily ambitious dream, but the trick is how do you break it down into manageable, achievable, interim steps that are meaningful and useful in their own right?

拥有一种疯狂的雄心勃勃的梦想很容易，但诀窍是如何将其分解为可管理、可实现、本身有意义且有用的临时步骤？

And so Virtual Cell, which is what I call the project of modeling a cell, I’ve had this idea of wanting to do that for maybe more like 25 years.

因此，Virtual Cell，也就是我所说的细胞建模项目，我想这样做的想法可能已经有 25 年了。

42:41 And I used to talk with Paul Nurse, who is a bit of a mentor of mine in biology.

我曾经和保罗·纳斯（Paul Nurse）交谈过，他可以说是我在生物学方面的导师。

He runs the founded the Crick Institute and won the Nobel Prize in 2001.

他创办了克里克研究所，并于 2001 年获得诺贝尔奖。

We’ve been talking about it since the nineties, and I used to come back to it every five years.

我们从九十年代起就一直在谈论这个问题，我过去每五年都会回顾一下这个问题。

It’s like, what would you need to model the full internals of a cell so that you could do experiments on the virtual cell and what those experiment in silico and those predictions would be useful for you to save you a lot of time in the wet lab.

就像，你需要什么来模拟细胞的完整内部结构，以便你可以在虚拟细胞上进行实验，以及这些计算机实验和这些预测对你有用，可以节省你在湿实验室中的大量时间。

That would be the dream.

那将是梦想。

43:13 Maybe you could a hundred x speed up experiments by doing most of it in silico the search in silico, and then you do the validation step in the wet lab.

也许您可以通过在计算机中进行大部分搜索（在计算机中进行搜索）来将实验速度加快一百倍，然后在湿实验室中进行验证步骤。

That’s the dream.

这就是梦想。

But maybe now, finally, so I was trying to build these components, AlphaFold being one, that would allow you eventually to model the full interaction, a full simulation of a cell, and I’d probably start with a yeast cell.

但也许现在，最后，我试图构建这些组件，AlphaFold 就是其中之一，它最终可以让你对完整的交互进行建模，对细胞进行完整的模拟，我可能会从酵母细胞开始。

And partly that’s what Paul Nurse studied because the yeast cell is like a full organism, that’s a single cell.

保罗·纳斯所研究的部分内容是因为酵母细胞就像一个完整的有机体，是一个单细胞。

So it’s the kind of simplest single cell organism.

所以它是一种最简单的单细胞生物。

And so it’s not just a cell, it’s a full organism.

所以它不仅仅是一个细胞，它是一个完整的有机体。

43:49 And yeast is very well understood.

酵母是很好理解的。

And so that would be a good candidate for a kind of full simulated model.

因此，这将是一种完整模拟模型的良好候选者。

Now AlphaFold is the solution to the kind of static picture of what does a 3D structure protein look like?

现在 AlphaFold 是 3D 结构蛋白质的静态图片的解决方案？

A static picture of it.

它的静态图片。

But we know that biology, all the interesting things happen with the dynamics, the interactions, and that’s what AlphaFold 3 is, the first step towards is modeling those interactions.

但我们知道，在生物学中，所有有趣的事情都是通过动力学、相互作用而发生的，这就是 AlphaFold 3 的本质，第一步就是对这些相互作用进行建模。

So first of all, pair wise proteins with proteins, proteins with RNA and DNA.

首先，将蛋白质与蛋白质、蛋白质与 RNA 和 DNA 配对。

But then the next step after that would be modeling maybe a whole pathway, maybe like the tour pathway that’s involved in cancer or something like this.

但接下来的下一步可能是对整个路径进行建模，也许像与癌症或类似的东西有关的旅行路径。

And then eventually you might be able to model a whole cell.

最终你也许能够对整个细胞进行建模。

Also, there’s another complexity here that stuff in a cell happens at different time scales.

此外，这里还有另一个复杂性，即细胞中的物质发生在不同的时间尺度。

44:31 Is that tricky?

这很棘手吗？

Protein folding is super fast.

蛋白质折叠速度超快。

I don’t know all the biological mechanisms, but some of them take a long time.

我不知道所有的生物学机制，但其中一些需要很长时间。

And so the levels of interaction has a different temporal scale that you have to be able to model.

因此，交互级别具有不同的时间尺度，您必须能够对其进行建模。

So that would be hard.

所以那会很难。

44:54 So you’d probably need several simulated systems that can interact at these different temporal dynamics, or at least maybe it’s like a hierarchical system so you can jump up or down the different temporal stages.

因此，您可能需要几个可以在这些不同的时间动态下交互的模拟系统，或者至少它就像一个分层系统，这样您就可以在不同的时间阶段上跳或下跳。

So can you avoid…

那么你能避免...

45:07 One of the challenges here is not avoid simulating, for example, the quantum mechanical aspects of any of this, right?

这里的挑战之一是不可避免地模拟其中任何一个的量子力学方面，对吧？

You want to not over model.

你不想过度建模。

You can skip ahead to just model the really high level things that get you a really good estimate of what’s going to happen.

您可以直接跳到对真正高级的事物进行建模，这可以让您对将要发生的情况有一个很好的估计。

Yes.

是的。

45:27 So you got to make a decision when you’re modeling any natural system, what is the cutoff level of the granularity that you’re going to model it to?

因此，在对任何自然系统进行建模时，您必须做出决定，您要对其进行建模的粒度的截止级别是多少？

And then it captures the dynamics that you’re interested in.

然后它会捕获您感兴趣的动态。

So probably for a cell I would hope that would be the protein level, and that one wouldn’t have to go down to the atomic level.

因此，对于细胞来说，我可能希望是蛋白质水平，而不必下降到原子水平。

So of course that’s where AlphaFold stock kicks in.

当然，这就是 AlphaFold 股票发挥作用的地方。

So that would be kind of the basis, and then you’d build these higher level simulations that take those as building blocks and then you get the emergent behavior.

所以这将是一种基础，然后你将构建这些更高级别的模拟，将它们作为构建块，然后你就会得到紧急行为。

I apologize for the pothead questions ahead of time, but do you think we’ll be able to simulate a model, the origin of life?

我为提前提出的问题表示歉意，但你认为我们能够模拟生命起源的模型吗？

46:00 So being able to simulate the first from non-living organisms, the birth of a living organism?

那么能够模拟第一个从非生物体，到生物体的诞生吗？

I think that’s one of course one of the deepest and most fascinating questions.

我认为这当然是最深刻、最有趣的问题之一。

46:19 I love that area of biology.

我喜欢生物学的那个领域。

There’s people, there’s a great book by Nick Lane, one of the top experts in this area called The Ten Great Inventions of Evolution.

有人，尼克·莱恩（Nick Lane）写了一本很棒的书，他是该领域的顶级专家之一，名为《进化的十大发明》。

I think it’s fantastic.

我认为这太棒了。

And it also speaks to what the great filters might be, prior or are they ahead of us?

它还说明了伟大的过滤器可能是什么，先于我们还是领先于我们？

I think they’re most likely in the past, if you read that book of how unlikely to go have any life at all.

如果你读过那本关于根本不可能拥有任何生活的书，我认为它们很可能已经过去了。

And then single cell to multi-cell seems an unbelievably big jump that took a billion years, I think on earth to do, right?

然后单细胞到多细胞似乎是一个令人难以置信的大跳跃，花了十亿年的时间，我想这在地球上是做不到的，对吗？

So it shows you how hard it was.

所以它告诉你这是多么困难。

Right?

正确的？

46:55 Bacteria were super happy for a very long time.

细菌在很长一段时间内都非常高兴。

For a very long time before they captured mitochondria somehow, right?

在他们以某种方式捕获线粒体之前很长一段时间，对吧？

46:56 I don’t see why not, why AI couldn’t help with that.

我不明白为什么不，为什么人工智能对此无能为力。

Some kind of simulation.

某种模拟。

Again, it’s a bit of a search process through a combinatorial space.

同样，这有点像通过组合空间进行搜索的过程。

Here’s all the chemical soup that you start with, the primordial soup, that maybe was on earth near these hot vents.

这是你开始使用的所有化学汤，原始汤，可能就在地球上这些热喷口附近。

Here’s some initial conditions.

这是一些初始条件。

Can you generate something that looks like a cell?

你能生成一些看起来像细胞的东西吗？

So perhaps that would be a next stage after the virtual cell project is well, how could something like that emerge from the chemical soup?

因此，也许这将是虚拟细胞项目顺利完成后的下一个阶段，这样的东西怎么可能从化学汤中出现呢？

Well, I would love it if there was a Move 37 for the origin of life.

好吧，如果生命起源有第 37 招，我会很高兴。

47:31 I think that’s one of the great mysteries.

我认为这是最大的谜团之一。

I think ultimately what we’ll figure out is their continuum.

我认为最终我们会弄清楚它们的连续体。

There’s no such thing as a line between non-living and living.

非生命与生命之间不存在界限。

But if we can make that rigorous.

但如果我们能够做到严格的话。

Yes.

是的。

That the very thing from the Big Bang to today has been the same process.

从大爆炸到今天的事情都是同样的过程。

47:45 If you can break down that wall that we’ve constructed in our minds of the actual origin from non-living to living, and it’s not a line that it’s a continuum that connects physics and chemistry and biology.

如果你能打破我们在头脑中建造的从非生命到生命的实际起源的那堵墙，它就不是一条线，它是连接物理、化学和生物学的连续体。

There’s no line.

没有线路。

I mean, this is my whole reason why I worked on AI and AGI my whole life, because I think it can be the ultimate tool to help us answer these kinds of questions.

我的意思是，这就是我一生致力于人工智能和通用人工智能的全部原因，因为我认为它可以成为帮助我们回答此类问题的终极工具。

48:03 And I don’t really understand why the average person doesn’t worry about this stuff more.

我真的不明白为什么普通人不更担心这些事情。

How can we not have a good definition of life and not living a non-living and the nature of time and let alone consciousness and gravity and all these things and quantum mechanics weirdness?

我们怎么能不能对生命有一个好的定义，而不是非生命和时间的本质，更不用说意识和重力以及所有这些东西和量子力学的怪异了？

It’s just to me, I’ve always had this sort of screaming at me in my face and it’s getting louder.

只是对我来说，我的脸上总是有人对我尖叫，而且声音越来越大。

It’s like, what is going on here?

就像，这里发生了什么事？

And I mean that in the deeper sense, the nature of reality, which has to be the ultimate question that would answer all of these things.

我的意思是，从更深层的意义上说，现实的本质，必须是回答所有这些问题的终极问题。

It’s sort of crazy if you think about it.

如果你仔细想想，这有点疯狂。

We can stare at each other and all these living things all the time.

我们可以一直凝视彼此以及所有这些生物。

We can inspect it microscopes and take it apart almost down to the atomic level.

我们可以用显微镜检查它，​​并将其拆解到几乎原子水平。

And yet we still can’t answer that clearly in a simple way.

然而我们仍然无法用简单的方式清楚地回答这个问题。

That question of how do you define living?

你如何定义生活这个问题？

It’s kind of amazing.

这真是太神奇了。

Yeah, living, you can kind of talk your way out of thinking about.

是的，活着，你可以通过谈论自己的方式来摆脱思考。

49:05 But consciousness, we have this very obviously subjective, conscious experience like we’re at the center of our own world and feels like something.

但是意识，我们有这种非常明显主观的、有意识的体验，就像我们处于自己世界的中心，感觉就像是某种东西。

And then how are you not screaming at the mystery of it all?

那么你怎么能不对这一切的奥秘尖叫呢？

I mean, but really humans have been contending with the mystery of the world around them for a long, long…

我的意思是，但实际上，人类长期以来一直在与周围世界的奥秘作斗争……

There’s a lot of mysteries like what’s up with the sun and the rain?

有很多谜团，比如太阳和雨是怎么回事？

What’s that about?

那是关于什么的？

And then last year we had a lot of rain, and this year we don’t have rain.

去年我们下了很多雨，今年我们没有下雨。

What did we do wrong?

我们做错了什么？

Humans have been asking that question for a long time.

人类长期以来一直在问这个问题。

Exactly.

确切地。

49:46 So I guess we’ve developed a lot of mechanisms to cope with these deep mysteries that we can’t fully, we can see, but we can’t fully understand and we have to just get on with daily life.

所以我想我们已经开发出了很多机制来应对这些我们无法完全理解、我们可以看到、但我们无法完全理解的深层奥秘，我们只能继续日常生活。

And we keep ourselves busy in a way.

我们在某种程度上让自己忙碌起来。

In a way, did we keep ourselves distracted?

在某种程度上，我们是否让自己分心了？

I mean, weather is one of the most important questions of human history.

我的意思是，天气是人类历史上最重要的问题之一。

50:01 We still, that’s the go-to small talk direction of the weather.

我们仍然认为，这是天气的闲聊方向。

Yes.

是的。

50:09 Especially in England.

尤其是在英格兰。

And then which is famously is an extremely difficult system to model.

众所周知，这是一个极其难以建模的系统。

50:11 And even that system, Google DeepMind has made progress on.

即使是这个系统，谷歌 DeepMind 也取得了进展。

Yes, we’ve created the best weather prediction systems in the world and they’re better than traditional fluid dynamics sort of systems that usually calculated on massive supercomputers takes days to calculate it.

是的，我们已经创建了世界上最好的天气预报系统，它们比传统的流体动力学系统更好，这些系统通常在大型超级计算机上需要几天的时间才能计算出来。

50:22 And we’ve managed to model a lot of the weather dynamics with neural network systems, with our WeatherNet system.

我们已经成功地利用神经网络系统和 WeatherNet 系统对许多天气动态进行了建模。

And again, it’s interesting that those kinds of dynamics can be modeled even though very complicated, almost bordering on chaotic systems in some cases.

有趣的是，即使这些动态非常复杂，在某些情况下几乎接近混沌系统，也可以对其进行建模。

50:50 A lot of the interesting aspects of that can be modeled by these neural network systems, including very recently we had cyclone prediction of where paths of hurricanes might go.

其中许多有趣的方面都可以通过这些神经网络系统进行建模，包括最近我们对飓风可能走向的路径进行了气旋预测。

Of course, super useful, super important for the world and it’s super important to do that very timely and very quickly and as well as accurately.

当然，这对世界来说非常有用、非常重要，而且非常及时、非常快速、准确地做到这一点非常重要。

And I think it’s very promising direction again, of simulating so that you can run forward predictions and simulations of very complicated real world systems.

我认为这又是一个非常有前途的方向，通过模拟，您可以对非常复杂的现实世界系统进行预测和模拟。

I should mention that I’ve gotten a chance in Texas to meet a community of folks called the Storm Chasers.

我应该提到的是，我有机会在德克萨斯州见到了一个名为“风暴追逐者”的社区。

51:18 And what’s really incredible about them, I need to talk to them more, is they’re extremely tech-savvy because what they have to do is they have to use models to predict where the storm is.

他们真正令人难以置信的是，我需要与他们更多地交谈，他们非常精通技术，因为他们要做的就是使用模型来预测风暴的位置。

So it’s this beautiful mix of crazy enough to go into the eye of the storm and in order to protect your life and predict where the extreme events are going to be, they have to have increasingly sophisticated models of weather.

因此，这是一种足够疯狂的美丽组合，足以进入风暴眼，为了保护您的生命并预测极端事件将在哪里发生，他们必须拥有越来越复杂的天气模型。

Yeah.

是的。

It is a beautiful balance of being in it as living organisms and the cutting edge of science.

这是生物体与科学前沿之间的完美平衡。

51:50 They actually might be using DeepMind systems.

他们实际上可能正在使用 DeepMind 系统。

Yeah.

是的。

52:01 But hopefully they are.

但希望他们是。

And I’d love to join them in one of those checks.

我很乐意和他们一起参加其中一张支票。

They look amazing.

他们看起来棒极了。

Right.

正确的。

That’s great to actually experience it one time.

能够实际体验一次真是太好了。

Exactly.

确切地。

52:07 And then also to experience the correct prediction where something will come and how it’s going to evolve.

然后还要体验正确的预测某事将在哪里发生以及它将如何演变。

It’s incredible.

太不可思议了。

You’ve estimated that we’ll have AGI by 2030, so there’s interesting questions around that.

您估计我们将在 2030 年实现 AGI，所以围绕这个问题存在一些有趣的问题。

How will we actually know that we got there and what may be the move quote, “Move 37” of AGI.

我们如何真正知道我们已经到达那里以及 AGI 的移动引用是什么，“移动 37”。

My estimate is sort of 50% chance by in the next five years, so by 2030 let’s say.

我的估计是在未来五年内，也就是到 2030 年，这种可能性有 50%。

52:33 So I think there’s a good chance that that could happen.

所以我认为这很有可能发生。

Part of it is what is your definition of AGI?

其中一部分是你对 AGI 的定义是什么？

Of course people arguing about that now and mind’s quite a high bar and always has been of can we match the cognitive functions that the brain has?

当然，人们现在争论的问题是，头脑的门槛相当高，而且一直如此，我们能否与大脑具有的认知功能相匹配？

So we know our brains are pretty much general Turing machines approximate, and of course we created incredible modern civilization with our minds.

所以我们知道我们的大脑与一般图灵机非常相似，当然我们用我们的思想创造了令人难以置信的现代文明。

So that also speaks to how general the brain is.

这也说明了大脑的通用性。

53:06 And for us to know we have a true AGI, we would have to make sure that it has all those capabilities.

为了让我们知道我们拥有真正的通用人工智能，我们必须确保它具有所有这些功能。

It isn’t kind of a jagged intelligence where some things, it’s really good at, like today’s systems, but other things it’s really flawed at.

它并不是一种参差不齐的智能，它在某些事情上确实很擅长，比如今天的系统，但在其他事情上它确实有缺陷。

And that’s what we currently have with today’s systems.

这就是我们目前的系统所拥有的。

They’re not consistent.

它们并不一致。

So you’d want that consistency of intelligence across the board.

所以你会希望情报全面一致。

53:28 And then we have some missing, I think, capabilities like the true invention capabilities and creativity that we were talking about earlier.

我认为，我们还缺少一些能力，比如我们之前讨论的真正的发明能力和创造力。

So you’d want to see those.

所以你想看看这些。

How you test that?

你如何测试？

I think you just test it.

我认为你只是测试一下。

One way to do it would be kind of brute force test of tens of thousands of cognitive tasks that we know that humans can do.

一种方法是对我们知道人类可以完成的数以万计的认知任务进行强力测试。

And maybe also make the system available to a few hundred of the world’s top experts, the Terence Taos of each subject area and give them a month or two and see if they can find an obvious flaw in the system.

也许还可以将该系统提供给数百名世界顶级专家（每个学科领域的特伦斯·陶斯），并给他们一两个月的时间，看看他们是否能找到系统中的明显缺陷。

And if they can’t, then I think you can be pretty confident we have a fully general system.

如果他们不能，那么我想你可以非常有信心我们拥有一个完全通用的系统。

Maybe to push back a little bit, it seems like humans are really incredible as the intelligence improves across all domains to take it for granted, like you mentioned, Terence Tao, these brilliant experts.

也许退一步来说，人类似乎真的很不可思议，因为所有领域的智力都在提高，认为这是理所当然的，就像你提到的，陶哲轩，这些才华横溢的专家。

54:11 They might quickly in a span of weeks, take for granted all the incredible things it can do and then focus in on, well, aha, right there.

他们可能会在几周内迅速认为它可以做的所有令人难以置信的事情是理所当然的，然后专注于，嗯，啊哈，就在那里。

I consider myself, first of all, human.

我首先认为自己是一个人。

I identify as human.

我认为自己是人类。

Some people listen to me talk and they’re like, “That guy is not good at talking the stuttering.” So even humans have obvious across domains limits, even just outside of calc, mathematics and physics and so on.

有些人听我说话，就会说：“这家伙说话不怎么会结巴。”因此，即使人类也有明显的跨领域限制，即使只是在计算、数学和物理等领域之外。

I wonder if it will take something like a Move 37, so on the positive side versus a barrage of 10,000 cognitive tasks where it’ll be one or two where it’s like, holy shit, this is special.

我想知道它是否需要像 Move 37 这样的东西，所以从积极的一面来看，与一万个认知任务相比，其中只有一两个会让人觉得，天啊，这很特别。

So I think that.

所以我认为。

55:14 Exactly.

确切地。

So I think there’s the sort of blanket testing to just make sure you’ve got the consistency.

所以我认为需要进行全面测试来确保一致性。

But I think there are the sort of lighthouse moments like the Move 37 that I would be looking for.

但我认为有像 Move 37 这样的灯塔时刻是我所寻找的。

So one would be inventing a new conjecture or a new hypothesis about physics like Einstein did.

因此，人们会像爱因斯坦那样发明一种关于物理学的新猜想或新假设。

55:33 So maybe you could even run the back test of that very rigorously, have a cut-off of 1900 and then give the system everything that was written up to 1900 and then see if it could come up with special relativity and general relativity, right?

所以也许你甚至可以非常严格地进行回溯测试，以 1900 为截止点，然后给系统提供 1900 之前写入的所有内容，然后看看它是否可以得出狭义相对论和广义相对论，对吗？

Like Einstein did.

就像爱因斯坦那样。

That would be an interesting test.

那将是一个有趣的测试。

Another one would be can it invent a game like Go?

另一个问题是它能发明像围棋这样的游戏吗？

Go not just come up with Move 37, a new strategy, but can it invent a game that’s as deep as aesthetically beautiful, as elegant as Go?

Go 不仅提出了第 37 步新策略，而且它能否发明一种像围棋一样有深度、美观、优雅的游戏？

And those are the sorts of things I would be looking out for.

这些都是我会关注的事情。

And probably a system being able to do several of those things for it to be very general, not just one domain.

一个系统可能能够做其中几件事，因为它是非常通用的，而不仅仅是一个领域。

And so I think that would be the signs at least that I would be looking for, that we’ve got a system that’s AGI level and then maybe to fill that out, you would also check their consistency, make sure there’s no holes in that system either.

所以我认为这至少是我要寻找的迹象，我们有一个 AGI 级别的系统，然后也许要填写它，你还要检查它们的一致性，确保该系统中也没有漏洞。

Yeah, something like a new conjecture or scientific discovery.

是的，就像一个新的猜想或科学发现。

56:27 That would be a cool feeling.

那会是一种很酷的感觉。

Yeah, that would be amazing.

是的，那太棒了。

56:32 So it’s not just helping us do that, but actually coming up with something brand new.

所以它不仅帮助我们做到这一点，而且实际上提出了一些全新的东西。

And you would be in the room for that.

你也会在场。

Absolutely.

绝对地。

It would be probably two or three months before announcing it.

大概要过两三个月才会公布。

56:40 And you would just be sitting there trying not to Tweet.

而你只是坐在那里，尽量不发推文。

Something like that.

类似的事情。

56:49 Exactly.

确切地。

It’s like what is this amazing new physics idea?

这就像这个令人惊奇的新物理想法是什么？

And then we would probably check it with world experts in that domain and validate it and go through its workings.

然后我们可能会与该领域的世界专家一起检查并验证它并检查其工作原理。

And I guess it would be explaining its workings too.

我想它也会解释它的工作原理。

Yeah.

是的。

It would be an amazing moment.

这将是一个美妙的时刻。

Do you worry that we as humans, even expert humans, like you might miss it?

您是否担心我们作为人类，甚至像您一样的专家，可能会错过它？

57:07 Might miss-

可能会错过——

Well, it may be pretty complicated.

嗯，这可能相当复杂。

57:12 So it could be, the analogy I give there is I don’t think it will be totally mysterious to the best human scientists, but it may be a bit like, for example in chess, if I was to talk to Garry Kasparov for Magnus Carlsen and play a game with them and they make a brilliant move, I might not be able to come up with that move.

所以，我给出的类比可能是，我不认为这对于最优秀的人类科学家来说是完全神秘的，但这可能有点像，例如在国际象棋中，如果我要与马格努斯·卡尔森的加里·卡斯帕罗夫交谈并与他们下棋，他们下了一个精彩的棋步，我可能无法想出那个棋步。

But they could explain why afterwards that move made sense.

但他们可以解释为什么后来这一举动是有意义的。

And we would be to understand it to some degree, not to the level they do, but if they were good at explaining, which is actually part of intelligence too, is being able to explain in a simple way that what you’re thinking about, I think that that will be very possible for the best human scientists.

我们会在某种程度上理解它，而不是达到他们的水平，但如果他们善于解释，这实际上也是智力的一部分，能够以简单的方式解释你正在思考的事情，我认为这对于最好的人类科学家来说是很有可能的。

But I wonder, maybe you can educate me on the side of Go, I wonder if there’s moves from Magnus or Garry where they at first will dismiss it as a bad move?

但我想知道，也许你可以在围棋方面教育我，我想知道马格努斯或加里是否有一些棋步，他们一开始会认为这是一个糟糕的棋步？

Yeah, sure, it could be.

是的，当然，可能是这样。

58:02 But then afterwards they’ll figure out with their intuition why this works.

但之后他们会凭直觉弄清楚为什么这是有效的。

And then empirically, the nice thing about games is, one of the great things about games is it’s a sort of scientific test.

然后从经验来看，游戏的好处是，游戏的伟大之处之一是它是一种科学测试。

Do you win the game or not win?

比赛你赢还是没赢？

And then that tells you, okay, that move in the end was good, that strategy was good.

然后这告诉你，好吧，最终的举动是好的，策略是好的。

And then you can go back and analyze that and explain even to yourself a little bit more why.

然后你可以回过头来分析这一点，甚至向自己解释一下原因。

Explore around it, and that’s how chess analysis and things like that works.

围绕它进行探索，这就是国际象棋分析之类的工作原理。

So perhaps that’s why my brain works like that because I’ve been doing that since I was four and it’s sort of hardcore training in that way.

所以也许这就是为什么我的大脑会这样工作，因为我从四岁起就一直这样做，这是一种硬核训练。

But even now when I generate code, there is this kind of nuanced, fascinating contention that’s happening where I might at first identify as a set of generated code is incorrect in some interesting nuanced ways.

但即使是现在，当我生成代码时，也会发生这种微妙的、令人着迷的争论，我可能首先会认为一组生成的代码在某些有趣的细微差别方面是不正确的。

58:39 But then I always have to ask the question, is there a deeper insight here that I’m the one who’s incorrect?

但我总是要问一个问题，是否有更深层次的洞察力表明我是错误的？

And that’s going to, as the systems get more and more intelligent, you’re going to have to contend with that.

随着系统变得越来越智能，你将不得不应对这个问题。

It’s like, is this a bug or a feature, what you just came up with?

就像，这是你刚刚想出的一个错误还是一个功能？

Yeah.

是的。

59:14 And they’re going to be pretty complicated to do, but of course it will be, you can imagine also AI systems that are producing that code or whatever that is, and then human programmers looking at it, but also not unaided with the help of AI tools as well.

做起来会非常复杂，但当然会，你也可以想象人工智能系统正在生成代码或其他任何东西，然后人类程序员查看它，但也不能没有人工智能工具的帮助。

So it’s going to be kind of an interesting, maybe different AI tools to the ones the monitoring tools are the ones that generated it.

因此，这将是一种有趣的、可能与生成它的监控工具不同的人工智能工具。

So if we look at that AGI system, sorry to bring it back up, but AlphaEvolve, it’s super cool.

因此，如果我们看看那个 AGI 系统，很抱歉将其重新启动，但是 AlphaEvolve，它非常酷。

59:36 So AlphaEvolve enables on the programming side, something like recursive self-improvement potentially.

因此，AlphaEvolve 在编程方面实现了类似递归自我改进的潜力。

If you can imagine what that AGI system, maybe not the first version, but a few versions beyond that, what does that actually look like?

如果你能想象那个 AGI 系统是什么，也许不是第一个版本，而是除此之外的几个版本，它实际上是什么样子？

Do you think it’ll be simple?

你认为这会很简单吗？

Do you think it’ll be something like a self-improving-

你认为这会是一种自我提升吗？

1:00:00 Like, do you think it’ll be simple?

比如说，你认为这会很简单吗？

Do you think it’ll be something like a self-improving program and a simple one?

你认为这会是一个自我完善的程序并且是一个简单的程序吗？

I mean potentially that’s possible.

我的意思是这有可能。

1:00:06 I would say I’m not sure it’s even desirable because that’s a kind of hard takeoff scenario.

我想说我不确定它是否可取，因为这是一种艰难的起飞场景。

But these current systems like Alpha Evolve, they have human in the loop deciding on various things, there’re separate hybrid systems that interact.

但目前的这些系统，比如 Alpha Evolve，它们让人类在循环中决定各种事情，有独立的混合系统进行交互。

1:00:22 One could imagine eventually doing that end to end.

人们可以想象最终会端到端地做到这一点。

I don’t see why that wouldn’t be possible, but right now I think the systems are not good enough to do that in terms of coming up with the architecture of the code.

我不明白为什么这是不可能的，但现在我认为系统在提出代码架构方面还不够好，无法做到这一点。

And again, it’s a little bit reconnected to this idea of coming up with a new conjectural hypothesis, how they’re good if you give them very specific instructions about what you’re trying to do, but if you give them a very vague high level instruction, that wouldn’t work currently.

再说一遍，这与提出一个新的猜想假设的想法重新联系起来，如果你给他们关于你想要做什么的非常具体的指示，他们会如何很好，但如果你给他们一个非常模糊的高级指示，那目前是行不通的。

And I think that’s related to this idea of invent a game as good as Go, right?

我认为这与发明一款像围棋一样好的游戏的想法有关，对吧？

1:00:55 Imagine that was the prompt.

想象一下这就是提示。

That’s pretty.

那很漂亮。

And so the current systems wouldn’t know I think what to do with that, how to narrow that down to something tractable.

因此，当前的系统不知道我想如何处理它，如何将其范围缩小到易于处理的范围。

And I think there’s similar, look, just make a better version of yourself.

我认为也有类似的地方，看看，只是让自己变得更好。

That’s too unconstrained.

这也太不受约束了吧。

But we’ve done it.

但我们已经做到了。

And as you know with AlphaVol, like things like faster matrix multiplication, so when you hone it down to very specific thing you want, it’s very good at incrementally improving that.

正如你所知道的 AlphaVol，就像更快的矩阵乘法之类的东西，所以当你把它磨练到你想要的非常具体的东西时，它非常擅长逐步改进。

1:01:21 But at the moment these are more incremental improvements, sort of small iterations.

但目前这些都是增量改进，有点小迭代。

Whereas if you wanted a big leap in understanding, you’d need a much larger advance.

然而，如果你想要在理解上有一个大的飞跃，你就需要更大的进步。

Yeah.

是的。

1:01:34 But it could also be sort of the pushback against hard takeoff scenario.

但这也可能是对硬起飞情景的阻碍。

It could be just a sequence of incremental improvements, like matrix multiplication.

它可能只是一系列渐进式改进，例如矩阵乘法。

It has to sit there for days thinking how to incrementally improve a thing and it does solve recursively.

它必须在那里呆上几天思考如何逐步改进某件事，并且它确实递归地解决了问题。

And as you do more and more improvement, it’ll slow down.

当你做越来越多的改进时，它会减慢。

1:01:55 So there be, the path to AGI won’t be like a gradual improvement over time.

因此，通向 AGI 的道路不会随着时间的推移而逐渐改进。

Yes.

是的。

1:02:03 If it was just incremental improvements, that’s how it would look.

如果只是渐进式改进，那就是这样的。

So the question is, could it come up with a new leap like the Transformers architecture?

那么问题来了，它能否像变形金刚架构一样，带来新的飞跃呢？

Could it have done that back in 2017 when we did it and Brain did it?

早在 2017 年，当我们和 Brain 都做到了的时候，它就能做到这一点吗？

And it’s not clear that these systems, something our AlphaVol wouldn’t be able to do, make such a big leap.

目前尚不清楚这些系统（我们的 AlphaVol 无法做到）是否会取得如此大的飞跃。

So for sure these systems are good.

所以可以肯定的是，这些系统都是好的。

We have systems I think that can do incremental hill climbing, and that’s a kind of bigger question about is that all that’s needed from here, or do we actually need one or two more big breakthroughs.

我认为我们有可以渐进式爬山的系统，这是一个更大的问题是，这里需要什么，或者我们实际上是否需要一两个更大的突破。

And can the same kind of systems provide the breakthroughs also?

同类系统也能提供突破吗？

1:02:34 So make it a bunch of S-curves like incremental improvement, but also every once in a while, leaps.

因此，让它成为一堆S曲线，就像渐进式改进一样，但也每隔一段时间，跳跃一次。

Yeah, I don’t think anyone has systems that can have shown, unequivocally those big leaps that we have a lot of systems that do the hill climbing of the S-curve that you’re currently on.

是的，我认为没有人拥有可以明确地展示那些巨大飞跃的系统，我们有很多系统可以实现您目前所处的 S 曲线爬坡。

And that would be the move 37 is a leap.

这就是第 37 步是一个飞跃。

Yeah, I think it would be a leap, something like that.

是的，我认为这将是一个飞跃，类似的事情。

Do you think the scaling laws are holding strong on the pre-training/post-training test time compute?

您认为缩放定律在训练前/训练后测试时间计算中是否有效？

1:03:01 Do you on the flip side of that, anticipate AI progress hitting a wall?

另一方面，您是否预计人工智能的进展会遇到困难？

We certainly feel there’s a lot more room just in the scaling.

我们当然认为在扩展方面还有更多的空间。

1:03:13 So actually all steps pre-training, post-training, and inference time.

所以实际上所有步骤都是训练前、训练后和推理时间。

So there’s sort of three scalings that are happening concurrently.

因此，三种扩展同时发生。

And again there, it’s about how innovative you can be and we pride ourselves on having the broadest and deepest research bench.

再说一遍，这关乎您的创新能力，我们为拥有最广泛和最深入的研究平台而感到自豪。

We have amazing, incredible researchers and people like Noam Shazir who came up with Transformers and Dave Silver who led the AlphaGo project and so on.

我们有令人惊叹的、令人难以置信的研究人员和像诺姆·沙齐尔（Noam Shazir）这样的人，他提出了变形金刚，戴夫·西尔弗（Dave Silver）领导了 AlphaGo 项目等等。

1:03:47 And that research base means that if some new breakthrough is required, like an AlphaGo or Transformers, I would back us to be the place that does that.

这个研究基地意味着，如果需要一些新的突破，比如 AlphaGo 或变形金刚，我会支持我们成为实现这一目标的地方。

So I’m actually quite like it when the terrain gets harder, right?

所以我实际上很喜欢地形变得更困难，对吧？

Because then it veers more from just engineering to true research, and research plus engineering, and that’s our sweet spot and I think that’s harder.

因为这样它就会更多地从工程转向真正的研究，以及研究加工程，这就是我们的最佳点，我认为这更难。

It’s harder to invent things than to fast follow.

发明事物比快速追随更难。

1:04:17 And so we don’t know, I would say it’s kind of 50/50 whether new things are needed or whether the scaling the existing stuff is going to be enough.

所以我们不知道，我想说这是 50/50 是否需要新东西或者扩展现有东西是否足够。

And so, in true kind of empirical fashion, we are pushing both of those as hard as possible.

因此，以真正的实证方式，我们正在尽可能努力地推动这两个目标。

The new blue sky, ideas and maybe about half our resources are on that.

新的蓝天、创意以及我们大约一半的资源都在于此。

And then scaling to the max, the current capabilities.

然后扩展到当前能力的最大值。

And we’re still seeing some fantastic progress on each different version of Gemini.

我们仍然看到 Gemini 的每个不同版本都取得了一些惊人的进展。

That’s interesting the way you put it in terms of the deep bench, that if progress towards AGI is more than just scaling compute, so the engineering side of the problem, and is more on the scientific side where there’s breakthroughs needed, then you feel confident DeepMind as well, Google DeepMind as well positioned to kick ass in that domain.

从深度工作台的角度来看，这很有趣，如果 AGI 的进展不仅仅是扩展计算，即问题的工程方面，而且更多的是需要突破的科学方面，那么您对 ​​DeepMind 也充满信心，谷歌 DeepMind 也有能力在该领域大放异彩。

Well, I mean if you look at the history of the last decade or 15 years, it’s been maybe, I don’t know, 80-90% of the breakthroughs that underpins modern AI field today was from originally Google Brain, Google Research and DeepMind.

嗯，我的意思是，如果你看看过去十年或 15 年的历史，我不知道，支撑当今现代人工智能领域的突破的 80-90% 可能最初来自 Google Brain、Google Research 和 DeepMind。

1:05:13 So yeah, I would back that to continue hopefully.

所以是的，我会支持这一点并希望继续下去。

So on the data side, are you concerned about running out of high quality data, especially high quality human data?

那么在数据方面，您是否担心高质量数据，尤其是高质量的人类数据用完？

I’m not very worried about that.

我对此不太担心。

1:05:37 Partly because I think there’s enough data, and it’s been proven to get the systems to be pretty good.

部分原因是我认为有足够的数据，并且已被证明可以使系统变得非常好。

And this goes back to simulations again.

这又回到了模拟。

Do you have enough data to make simulations, so that you can create more synthetic data that are from the right distribution?

您是否有足够的数据来进行模拟，以便您可以创建更多来自正确分布的合成数据？

Obviously that’s the key.

显然这是关键。

So you need enough real-world data in order to be able to create those kinds of data generators, and I think that we’re at that step at the moment.

因此，您需要足够的现实世界数据才能创建此类数据生成器，我认为我们目前正处于这一步。

Yeah, you’ve done a lot of incredible stuff on the side of science and biology, doing a lot with not so much data.

是的，你在科学和生物学方面做了很多令人难以置信的事情，但用的数据并不多。

Yeah.

是的。

I mean it’s still a lot of data, but I guess enough to-

我的意思是它仍然有很多数据，但我想足以-

To get that going.

为了让这一切顺利进行。

1:06:15 Exactly.

确切地。

Exactly

正是如此

Yeah.

是的。

1:06:18 How crucial is the scaling of compute to building AGI?

计算扩展对于构建 AGI 有多重要？

That’s an engineering question.

这是一个工程问题。

It’s almost a geopolitical question because it also integrated into that is supply chains and energy.

这几乎是一个地缘政治问题，因为它也融入了供应链和能源。

A thing that you care a lot about, which is potentially fusion.

你非常关心的一件事，就是潜在的融合。

So innovating on the side of energy also.

因此，能源方面也进行了创新。

Do you think we’re going to keep scaling compute?

您认为我们会继续扩展计算吗？

I think so, for several reasons.

我这么认为，有几个原因。

1:06:42 I think compute, there’s the amount of compute you have for training, often it needs to be co-located, so actually even bandwidth constraints between data centers can affect that.

我认为计算，有用于训练的计算量，通常需要位于同一地点，所以实际上，即使数据中心之间的带宽限制也会影响这一点。

So there’s additional constraints even there and that’s important for training, obviously the largest models you can, but there’s also because now AI systems are in products and being used by billions of people around the world, you need a ton of inference compute now.

因此，即使在那里也存在额外的限制，这对于训练很重要，显然是你可以训练的最大模型，但也因为现在人工智能系统已经进入产品并被全世界数十亿人使用，所以你现在需要大量的推理计算。

1:07:10 And then on top of that there’s the thinking systems, the new paradigm of the last year that where they get smarter, the longer amount of inference time you give them at test time.

除此之外还有思维系统，去年的新范式，它们变得更聪明，你在测试时给它们的推理时间就越长。

So all of those things need a lot of compute and I don’t really see that slowing down, and as AI systems become better, they’ll become more useful and there’ll be more demand for them.

因此，所有这些事情都需要大量的计算，而且我并没有真正看到这种速度会放缓，而且随着人工智能系统变得更好，它们将变得更有用，对它们的需求也会更大。

So both from the training side, the training side actually is only just one part of that.

所以无论是从培训方面来说，培训方面实际上只是其中的一部分。

It may even become the smaller part of what’s needed in the overall compute that’s required.

它甚至可能成为整个计算所需的较小部分。

Yeah, that’s one sort of almost meme-y kind of thing, which is the success in the incredible aspects of VL3.

是的，这就是一种近乎迷因的事情，这就是 VL3 令人难以置信的成功。

1:07:42 People kind of make fun of the more successful it becomes, the servers are sweating.

人们有点取笑它变得越成功，服务器就流汗了。

Yes.

是的。

The inference.

推论。

Yeah, yeah, exactly.

是的，是的，完全正确。

1:07:57 We did a little video of the servers frying eggs and things.

我们制作了一段服务器煎鸡蛋之类的小视频。

That’s right.

这是正确的。

And we are going to have to figure out how to do that.

我们必须弄清楚如何做到这一点。

There’s a lot of interesting hardware innovations that we do as we have our own TPU line and we’re looking at inference-only things, inference-only chips and how we can make those more efficient.

我们做了很多有趣的硬件创新，因为我们拥有自己的 TPU 系列，并且我们正在研究仅推理的东西、仅推理的芯片以及如何提高这些创新的效率。

1:08:15 We’re also very interested in building AI systems and we have done the help with energy usage, so help data center energy like for the cooling systems be efficient, grid optimization, and then eventually things like helping with plasma-containment fusion reactors.

我们也对构建人工智能系统非常感兴趣，并且我们已经在能源使用方面提供了帮助，因此帮助数据中心能源（例如冷却系统）变得高效、电网优化，然后最终帮助等离子体包容聚变反应堆。

We’ve done lots of work on that with Commonwealth Fusion, and also one could imagine reactor design.

我们已经通过 Commonwealth Fusion 做了很多这方面的工作，人们也可以想象反应堆的设计。

1:08:38 And then material design I think is one of the most exciting.

我认为材料设计是最令人兴奋的之一。

New types of solar material, solar panel material room temperature superconductors has always been on my list of dream breakthroughs, and optimal batteries.

新型太阳能材料、太阳能电池板材料、室温超导体一直是我梦想突破的清单，也是最佳电池。

And I think a solution to any one of those things would be absolutely revolutionary for climate and energy usage.

我认为解决其中任何一个问题对于气候和能源使用来说绝对是革命性的。

And we’re probably close, and again in the next five years to having AI systems that can materially help with those problems.

我们可能会在未来五年内再次接近拥有能够为解决这些问题提供实质性帮助的人工智能系统。

If you were to bet, sorry for the ridiculous question, but what is the main source of energy in 20, 30, 40 years.

如果你要打赌，抱歉问了这个可笑的问题，但是20年、30年、40年的主要能源来源是什么？

1:09:05 Do you think it’s going to be nuclear fusion?

你认为这会是核聚变吗？

I think fusion and solar are the two that I would bet on.

我认为聚变和太阳能是我押注的两个。

1:09:15 Solar, I mean it’s the fusion reactor in the sky of course, and I think really the problem there is batteries and transmission.

太阳能，我的意思是它当然是天空中的聚变反应堆，我认为真正的问题在于电池和传输。

So as well as more efficient, more and more efficient solar material perhaps eventually in space, these kind of Dyson Sphere type ideas.

因此，以及更高效、越来越高效的太阳能材料也许最终会出现在太空中，这些戴森球类型的想法。

1:09:36 And fusion I think is definitely doable, it seems, if we have the right design of reactor and we can control the plasma and fast enough and so on, and I think both of those things will actually get solved.

我认为聚变绝对是可行的，如果我们有正确的反应堆设计，我们可以控制等离子体并足够快等等，我认为这两件事实际上都会得到解决。

So we’ll probably have at least those are probably the two primary sources of renewable, clean, almost free or perhaps free energy.

因此，我们可能至少会拥有可再生、清洁、几乎免费或可能免费能源的两种主要来源。

What a time to be alive.

活着是多么美好的时光啊。

1:09:58 If I traveled into the future with you a hundred years from now, how much would you be surprised if we’ve passed a type one Kardashev scale civilization?

如果我和你一起穿越到一百年后的未来，如果我们已经过了卡尔达肖夫级文明，你会感到多么惊讶？

I would not be that surprised if it was a hundred-year timescale from here.

如果时间跨度是一百年，我不会感到惊讶。

1:10:11 I mean I think it’s pretty clear if we crack the energy problems in one of the ways we’ve just discussed or very efficient solar, then if energy is kind of free and renewable and clean, then that solves a whole bunch of other problems.

我的意思是，我认为很明显，如果我们以我们刚才讨论的方式之一解决能源问题，或者非常高效的太阳能，那么如果能源是免费的、可再生的和清洁的，那么这就解决了一大堆其他问题。

1:10:32 So for example, the water access problem goes away because you can just use desalination.

例如，用水问题就消失了，因为你可以使用海水淡化。

We have the technology, it’s just too expensive.

我们有技术，只是太贵了。

So only fairly wealthy countries like Singapore and Israel and so on actually use it.

所以只有新加坡和以色列等相当富裕的国家才会真正使用它。

But if it was cheap, then all countries that have a coast could, but also you’d have unlimited rocket fuel.

但如果它很便宜，那么所有拥有海岸的国家都可以，而且你还会拥有无限的火箭燃料。

You could just separate seawater out into hydrogen and oxygen using energy and that’s rocket fuel.

你可以利用能量将海水分离成氢气和氧气，这就是火箭燃料。

1:10:57 So combined with Elon’s, amazing self landing rockets, then it could be you sort of like a bus service to space.

因此，与埃隆令人惊叹的自着陆火箭相结合，那么它可能有点像通往太空的巴士服务。

So that opens up incredible new resources and domains.

这样就开辟了令人难以置信的新资源和领域。

Asteroid mining I think will become a thing, and maximum human flourishing to the stars.

我认为小行星采矿将成​​为一件事，人类将在星星上实现最大程度的繁荣。

That’s what I dream about as well is like Carl Sagan’s sort of idea of bringing consciousness to the universe, waking up the universe.

这也是我的梦想，就像卡尔·萨根的那种将意识带入宇宙、唤醒宇宙的想法。

And I think human civilization will do that in the full sense of time if we get AI right, and crack some of these problems with it.

我认为，如果我们正确使用人工智能，并用它解决其中一些问题，人类文明将在完全意义上做到这一点。

Yeah, I wonder what it would look like if you’re just a tourist flying through space.

是的，我想知道如果你只是一个穿越太空的游客会是什么样子。

1:11:30 You would probably notice earth because if you solve the energy problem, you would see a lot of space rockets probably.

你可能会注意到地球，因为如果你解决了能源问题，你可能会看到很多太空火箭。

So it would be traffic here in London, but in space.

因此，这将是伦敦这里的交通，但在太空中。

Yes, exactly.

是的，完全正确。

It’s just a lot of rockets.

这只是很多火箭。

1:11:46 And then you would probably see floating in space, some kind of source of energy like solar potentially.

然后你可能会看到漂浮在太空中的某种能源，例如太阳能。

So earth would just look more on the surface, more technological.

所以地球表面上看起来会更加科技化。

And then you would use the power of that energy then to preserve the natural…

然后你会利用这种能量的力量来保护自然……

Yes.

是的。

Like the rainforest and all that kind of stuff.

就像热带雨林和所有类似的东西。

Exactly.

确切地。

1:12:07 Because for the first time in human history we wouldn’t be resource constrained.

因为在人类历史上，我们第一次不再受到资源限制。

And I think that could be amazing new era for humanity where it’s not zero-sum, right?

我认为这对人类来说可能是一个令人惊奇的新时代，它不是零和的，对吗？

I have this land, you don’t have it.

这片土地我有，你没有。

Or if the tigers have their forest, then the local villages can’t, what are they going to use?

或者说，如果老虎有森林，当地村庄却没有，那他们拿什么去利用呢？

I think that this will help a lot.

我认为这会有很大帮助。

No, it won’t solve all problems because there’s still other human foibles that will still exist, but it will at least remove one, I think one of the big vectors, which is scarcity of resources, including land and more materials and energy.

不，它并不能解决所有问题，因为人类的其他弱点仍然存在，但它至少会消除一个，我认为是最大的矢量之一，那就是资源的稀缺，包括土地和更多的材料和能源。

1:12:45 And we should be sometimes call it another call about this kind of radical abundance era, where there’s plenty of resources to go around.

有时我们应该称其为关于这种激进的富足时代的另一种召唤，那里有大量的资源可供使用。

Of course the next big question is making sure that that’s fairly, shared fairly and everyone in society benefits from that.

当然，下一个大问题是确保公平、公平地分享，并且社会中的每个人都能从中受益。

So there is something about human nature where I go, its like Borat, like my neighbor.

所以我所到之处都有一些关于人性的东西，就像波拉特，就像我的邻居。

1:13:01 You start trouble.

你就开始惹麻烦了。

We do start conflicts and that’s why games throughout, as I’m learning actually more and more, even in ancient history, serve the purpose of pushing people away from war, actually hot war.

我们确实会引发冲突，这就是为什么游戏始终是为了让人们远离战争，实际上是热战，我实际上越来越多地了解到，即使在古代历史中也是如此。

So maybe we can figure out increasingly sophisticated video games that pull us, they give us that…

所以也许我们可以找出越来越复杂的电子游戏来吸引我们，它们给我们带来……

Scratch the itch of conflict, whatever that is, but us, the human nature.

挠痒痒的冲突，无论是什么，但我们，人性。

Like…

就像……

1:13:38 Yeah.

是的。

And then avoid the actual hot wars that would come with increasingly sophisticated technologies because we’re now, we’ve long passed the stage where the weapons we’re able to create can actually just destroy all of human civilization.

然后避免日益复杂的技术带来的实际热战，因为我们现在已经过了我们能够制造的武器实际上可以摧毁所有人类文明的阶段。

So that’s no longer a great way to start with your neighbor.

因此，这不再是从邻居开始的好方法。

It’s better to play a game of chess.

最好还是下一盘象棋。

Or football.

或者足球。

Or football.

或者足球。

1:14:03 Yeah.

是的。

And I think that’s what my modern sport is and I love football watching it and I just feel like, and I used to play it a lot as well, and it’s very visceral in its tribal, and I think it does channel a lot of those energies into which I think is a kind of human need to belong to some group, but into a fun way, a healthy way and not destructive way, kind of constructive thing.

我认为这就是我的现代运动，我喜欢看足球，我只是觉得，我也经常玩它，它的部落是发自内心的，我认为它确实输送了很多能量，我认为这是人类属于某个群体的需要，但以一种有趣的方式，一种健康的方式而不是破坏性的方式，一种建设性的方式。

1:14:33 And I think going back to games again is I think they’re originally why they’re so great as well for kids to play things like chess is they’re great little microcosm simulations of the world.

我认为再次回到游戏，我认为它们最初是为什么它们对孩子们玩国际象棋之类的东西如此伟大的原因是它们是世界的伟大的小微观模拟。

They’re simulations of the world too.

它们也是对世界的模拟。

They’re simplified versions of some real world situation, whether it’s poker or Go or chess, different aspects or diplomacy, different of the real world.

它们是一些现实世界情况的简化版本，无论是扑克、围棋还是国际象棋，不同的方面或外交，与现实世界不同。

1:14:53 And it allows you to practice at them too, because how many times do you get to practice a massive decision moment in your life?

它也可以让你练习它们，因为你一生中有多少次可以练习重大决策时刻？

What job to take, what university to go to?

做什么工作，去哪所大学？

You get maybe, I don’t know, a dozen or so key decisions one has to make and you’ve got to make those as best as you can.

我不知道，你可能需要做出十几个关键决定，而你必须尽你所能做出最好的决定。

And games is a kind of safe environment, repeatable environment where you can get better at your decision-making process, and it maybe has this additional benefit of channeling some energies into more creative and constructive pursuits.

游戏是一种安全的、可重复的环境，你可以在其中更好地制定决策，而且它可能还具有将一些精力引导到更具创造性和建设性的追求上的额外好处。

Well I think it’s also really important to practice losing and winning.

嗯，我认为练习输和赢也非常重要。

Right.

正确的。

Losing is a really, that’s why I love games.

失败确实是一种痛苦，这就是我热爱游戏的原因。

1:15:29 That’s why I love even things like Brazilian Jiu-Jitsu where you can get your kicked in a safe environment over and over.

这就是为什么我什至喜欢巴西柔术这样的东西，在那里你可以在安全的环境中一遍又一遍地踢你的拳。

It reminds you about physics, about the way the world works about sometimes you lose, sometimes you win, you can still be friends with everybody.

它让你想起物理学，想起世界的运作方式，有时你输了，有时你赢了，你仍然可以和每个人成为朋友。

But that feeling of losing, I mean it’s a weird one for us humans to really make sense of.

但那种失落的感觉，我的意思是，对于我们人类来说，真正理解这种感觉很奇怪。

That’s just part of life.

这只是生活的一部分。

That is a fundamental part of life is losing.

这是生命的一个基本部分正在失去。

And I think the martial arts as I understand it, but also in things like light chess is at least the way I took it’s a lot to do with self-improvement, self-knowledge.

我认为我所理解的武术，以及像轻棋这样的东西，至少是我所采取的方式，这与自我完善、自我认识有很大关系。

1:16:00 That, okay, so I did this thing.

那好吧，所以我做了这件事。

It’s not about really beating the other person, it’s about maximizing your own potential.

并不是要真正打败别人，而是要最大限度地发挥自己的潜力。

1:16:16 If you do it in a healthy way, you learn to use victory and losses in a way.

如果你以健康的方式去做，你就会学会以某种方式利用胜利和失败。

Don’t get carried away with victory and think you’re just the best in the world.

不要因胜利而得意忘形，认为自己是世界上最好的。

And the losses keep you humble, and always knowing there’s always something more to learn.

失败让你保持谦虚，并且始终知道总有更多东西需要学习。

There’s always a bigger expert that you can mentor you.

总有一位更优秀的专家可以指导您。

I think you learn that I’m pretty sure in martial arts.

我想你知道我对武术很有把握。

1:16:35 And I think that’s also the way that at least I was trained in chess.

我认为至少我接受国际象棋训练的方式也是如此。

And so, in the same way, and it can be very hardcore and very important and of course you want to win, but you also need to learn how to deal with setbacks in a healthy way, and wire that feeling that you have when you lose something into a constructive thing of, next time I’m going to improve this or get better at this.

所以，以同样的方式，它可以是非常核心和非常重要的，当然你想赢，但你也需要学习如何以健康的方式处理挫折，并将当你失去某些东西时的感觉转化为建设性的东西，下次我要改进它或在这方面做得更好。

There is something that’s a source of happiness, a source of meaning that improvements that…

有一些东西是幸福的源泉，是意义的源泉，可以改善……

1:16:57 It’s not about the winning or losing.

这与输赢无关。

Yes, the mastery.

是的，精通。

1:17:04 There’s nothing more satisfying in a way.

从某种程度上来说，没有什么比这更令人满意的了。

It’s like, oh wow, this thing I couldn’t do before.

就像，哦哇，这件事我以前做不到。

Now I can.

现在我可以了。

And again, games and physical sports and mental sports, their ways of measuring their beautiful, because you can measure that progress.

再说一次，游戏、体育运动和心理运动，是他们衡量美丽的方式，因为你可以衡量进步。

Yeah, there’s something about I guess why I love role-playing games, like the number go up of on the skill tree, literally that is a source of meaning for us humans, whatever our-

是的，我想我为什么喜欢角色扮演游戏，比如技能树上的数字，从字面上看，这对我们人类来说是意义的来源，无论我们 -

Yeah, we’re quite addicted to this sort of, these numbers going up.

是的，我们非常沉迷于这种数字的上升。

1:17:29 And maybe that’s why we made games like that because obviously that is something we’re hill climbing systems ourselves, right?

也许这就是我们制作这样的游戏的原因，因为显然这就是我们自己的爬山系统，对吧？

Yeah.

是的。

1:17:42 It would be quite sad if we didn’t have any mechanism-

如果我们没有任何机制那就太可悲了——

Color belts, we do this everywhere, where we just have thing that…

色带，我们到处都这样做，我们只是有这样的东西……

And I don’t want to dismiss that.

我不想否认这一点。

1:17:51 There is a source of deep meaning across humans.

人类之间存在着深刻的意义之源。

1:17:55 So one of the incredible stories on the business, on the leadership side is what Google has done over the past year.

因此，在商业和领导力方面，最令人难以置信的故事之一就是谷歌在过去一年中所做的事情。

So I think it’s fair to say that Google was losing on the LLM product side a year ago with Gemini 1.5 And now it’s winning, which…

所以我认为可以公平地说，谷歌一年前在 Gemini 1.5 的 LLM 产品方面输了，但现在它赢了，这……

I’m Joe Biden.

我是乔·拜登。

And you took the helm and you led this effort.

你掌舵并领导了这项工作。

What did it take to go from let’s say quote-unquote losing to quote-unquote winning, in the span of a year?

在一年的时间里，从惨败到胜利，经历了什么？

Yeah, well firstly it’s absolutely incredible team that we have led by Corey and Jeff Dean and Oriole and the amazing team we have on Gemini.

是的，首先，我们由科里（Corey）、杰夫·迪恩（Jeff Dean）和金莺（Oriole）领导的绝对令人难以置信的团队，以及我们在双子座上拥有的令人惊叹的团队。

1:18:22 Absolutely.

绝对地。

So you can’t do it without the best talent.

因此，如果没有最优秀的人才，你就无法做到这一点。

And of course we have a lot of great compute as well.

当然，我们也有很多出色的计算能力。

But then it’s the research culture we’ve created and basically coming together both different groups in Google that was Google Brain, World-class team, and then the old DeepMind, and pulling together all the best people and the best ideas and gathering around to make the absolute greater system we could.

但我们创造了研究文化，基本上将谷歌的两个不同团队聚集在一起，即谷歌大脑、世界一流的团队，然后是旧的 DeepMind，汇集所有最优秀的人才和最好的想法，并聚集在一起，以打造我们可以做到的绝对更伟大的系统。

1:18:59 And it was been hard, but we’re all very competitive and we love research.

这很艰难，但我们都非常有竞争力，而且我们热爱研究。

This is so fun to do, and it’s great to see our trajectory.

这样做非常有趣，很高兴看到我们的发展轨迹。

It wasn’t a given, but we’re very pleased with where we are and the rate of progress is the most important thing.

这不是理所当然的，但我们对目前的情况非常满意，进展速度是最重要的。

So if you look at where we’ve come to from two years ago to one year ago to now, I think we call it relentless progress.

因此，如果你看看我们从两年前到一年前到现在所取得的进展，我认为我们称之为不懈的进步。

Along with relentless shipping of that progress is being very successful and it’s unbelievably competitive, the whole space, the whole AI space, with some of the greatest entrepreneurs and leaders and companies in the world, all competing now because everyone’s realized how important AI is.

随着这一进步的不断推进，非常成功，并且竞争非常激烈，整个领域、整个人工智能领域，以及世界上一些最伟大的企业家、领导者和公司，现在都在竞争，因为每个人都意识到人工智能的重要性。

And it’s very been pleasing for us to see that progress.

我们很高兴看到这一进展。

Google’s a gigantic company.

谷歌是一家大公司。

1:19:47 Can you speak to the natural things that happen in that case is the bureaucracy that emerges?

您能谈谈在那种情况下发生的自然事情就是出现的官僚主义吗？

You want to be careful the natural, there’s meetings and there’s managers and that.

你自然要小心，有会议，有经理等等。

What are some of the challenges from a leadership perspective, breaking through that in order to, like you said, ship?

从领导力的角度来看，要突破这些挑战以便像您所说的那样进行交付，有哪些挑战？

Like the number of products, Gemini related products that has been shipped over the past years is insane.

与产品数量一样，过去几年 Gemini 相关产品的出货量也是疯狂的。

Right?

正确的？

1:20:14 Yeah, exactly.

是的，完全正确。

That’s what relentlessness looks like.

这就是无情的样子。

I think it’s a question of any big company ends up having a lot of layers of management and things like that is sort of the nature of how it works.

我认为这是一个问题，任何大公司最终都会有很多管理层，类似的事情就是它运作的本质。

But I still operate and I was always operating with old DeepMind as a start-up still.

但我仍然在运营，而且我一直在以初创公司的形式运营旧的 DeepMind。

A large one, but still as a start-up.

规模很大，但仍是一家初创公司。

1:20:37 And that’s what we still act like today with Google DeepMind.

这就是我们今天对 Google DeepMind 的做法。

And acting with decisiveness and the energy that you get from the best smaller organizations.

并以从最好的小型组织那里获得的果断行动和活力。

And we try to get the best of both worlds where we have this incredible, billions of users surfaces and credible products that we can power up with our AI and our research and that’s amazing and that’s very few places in the world you can get that, do incredible world-class research on the one hand and then plug it in and improve billions of people’s lives the next day.

我们试图充分利用这两个世界，我们拥有令人难以置信的数十亿用户界面和可靠的产品，我们可以通过我们的人工智能和我们的研究来增强这些产品，这真是太棒了，世界上很少有地方可以做到这一点，一方面进行令人难以置信的世界级研究，然后在第二天将其投入使用并改善数十亿人的生活。

That’s a pretty amazing combination.

这是一个非常惊人的组合。

1:21:10 And we’re continually fighting and cutting away bureaucracy to allow the research culture and the relentless shipping culture to flourish.

我们不断地与官僚作风作斗争，让研究文化和不懈的航运文化蓬勃发展。

And I think we’ve got a pretty good balance, whilst being responsible with it, as you have to be as a large company and also with a number of huge product surfaces that we have.

我认为我们在对此负责的同时取得了很好的平衡，因为你必须是一家大公司，并且我们拥有许多巨大的产品表面。

So a funny thing you mentioned about the surface with the billion, I had a conversation with a guy named, brilliant guy here at the British Museum, called Irvin Finkel.

你提到的关于十亿表面的有趣的事情是，我和大英博物馆的一位名叫欧文·芬克尔（Irvin Finkel）的才华横溢的人进行了交谈。

1:21:30 He’s a world expert at cuneiforms, which is a ancient writing on tablets and he doesn’t know about ChatGPT or Gemini, he doesn’t even know about AI, but this first encounter with this AI is AI mode on Google.

他是楔形文字的世界专家，楔形文字是平板电脑上的一种古老文字，他不知道 ChatGPT 或 Gemini，他甚至不知道人工智能，但第一次接触这种人工智能是谷歌的人工智能模式。

Yes.

是的。

He’s like, is that what you’re talking about, this AI mode?

他说，这就是你所说的人工智能模式吗？

1:21:58 And it’s just a reminder that there’s a large part of the world that doesn’t know about this AI thing.

这只是提醒我们，世界上有很大一部分人不了解人工智能这个东西。

Yeah, I know.

是的，我知道。

1:22:08 It’s funny.

这很有趣。

If you live on X and Twitter and I mean it’s sort of at least my feed, it’s all AI.

如果你生活在 X 和 Twitter 上，我的意思是，这至少是我的动态，那都是人工智能。

And there’s certain places where in the valley and certain pockets where everyone’s just, all they’re thinking about is AI, but a lot of the normal world hasn’t come across it yet.

在山谷的某些地方和某些地方，每个人都只是思考人工智能，但很多正常世界还没有遇到过它。

And that’s a great responsibility to their first interaction.

这对他们的第一次互动来说是一个巨大的责任。

1:22:24 The grand scale of the rural, India or anywhere across the world you get to…

印度或世界任何地方的乡村规模宏大……

And we want it to be as good as possible and in a lot of cases it’s just under the hood powering, making something like maps or search work better.

我们希望它尽可能好，在很多情况下，它只是在幕后提供动力，使地图或搜索等功能更好地工作。

1:22:34 And ideally for a lot of those people should just be seamless.

理想情况下，对于很多人来说应该是无缝的。

It’s just new technology that makes their lives more productive and helps them.

新技术可以让他们的生活更加富有成效并为他们提供帮助。

A bunch of folks on the Gemini product and engineering teams spoken extremely highly of you on another dimension, that I almost didn’t even expect.

Gemini 产品和工程团队的一群人在另一个层面对你给予了极高的评价，这是我几乎没有想到的。

1:22:50 I kind of think of you as the deep scientists and caring about these big research scientific questions.

我认为你们是深层次的科学家，关心这些重大的研究科学问题。

But they also said you’re a great product guy, like how to create a thing that a lot of people would use and enjoy using.

但他们也说你是一个很棒的产品人员，喜欢如何创造一个很多人会使用并喜欢使用的东西。

So can you maybe speak to what it takes to create a AI based product that a lot of people enjoy using?

那么您能否谈谈如何创建一个很多人喜欢使用的基于人工智能的产品？

Yeah.

是的。

1:23:18 Well, I mean, again, that comes back from my game design days where I used to design games for millions of gamers.

嗯，我的意思是，这又是从我的游戏设计时代开始的，当时我为数百万玩家设计游戏。

People would forget about that.

人们会忘记这一点。

I’ve had experience with cutting edge technology in product that is how games was in the nineties.

我在产品中拥有尖端技术的经验，就像九十年代的游戏一样。

1:23:31 And so I love actually the combination of cutting edge research and then being applied in a product and to power a new experience.

因此，我实际上喜欢前沿研究的结合，然后将其应用到产品中并为新体验提供动力。

And so, I think it’s the same skill really of imagining what it would be like to use it viscerally, and having good taste coming back to earlier.

因此，我认为这与发自内心地想象使用它会是什么样子以及早先拥有良好的品味实际上是相同的技能。

The same thing that’s useful in science, I think can also be useful in product design.

我认为在科学中有用的东西在产品设计中也很有用。

1:23:57 And I’ve just had a very, always been a sort of multidisciplinary person, so I don’t see the boundaries really between arts and sciences, or product and research.

我有一个非常非常多学科的人，所以我看不到艺术和科学、产品和研究之间的真正界限。

It’s a continuum for me.

这对我来说是一个连续体。

I like working on products that are cutting edge.

我喜欢研究前沿的产品。

I wouldn’t be able to have cutting edge technology under the hood.

我无法拥有最先进的技术。

I wouldn’t be excited about them if they were just run-of-the-mill products.

如果它们只是普通产品，我不会对它们感到兴奋。

It requires this invention, creativity, cap capability.

它需要这种发明、创造力、盖帽能力。

What are some specific things you learned about when you, even on the LLM side, you’re interacting with Gemini?

当你（即使是法学硕士）与双子座互动时，你学到了哪些具体的事情？

1:24:23 This doesn’t feel like, the layout, the interface, maybe the trade-off between the latency, how to present to the user, how long to wait and how that waiting is shown or the reason capabilities.

这感觉不像布局，界面，可能是延迟之间的权衡，如何呈现给用户，等待多长时间以及如何显示等待或原因功能。

There are some interesting things because like you said, it’s the very cutting edge.

有一些有趣的事情，因为就像你说的，这是非常前沿的。

We don’t know how to present it correctly.

我们不知道如何正确地呈现它。

So is there some specific things you’ve learned?

那么你学到了一些具体的东西吗？

I mean it’s such a false evolving space, evaluating this all the time, but where we are today is that you want to continually simplify things, whether that’s the interface or what you build on top of the model, you kind of want to get out of the way of the model.

我的意思是，这是一个错误的不断发展的空间，一直在评估这一点，但我们今天的情况是，你想要不断简化事物，无论是界面还是在模型之上构建的内容，你都想摆脱模型的束缚。

1:24:55 The model train is coming down the track and it’s improving unbelievably fast.

火车模型正在沿着轨道行驶，并且改进速度快得令人难以置信。

This relentless progress we talked about earlier.

我们之前谈到过这种持续不断的进步。

1:25:17 You look at 2.5 versus 1.5 and it’s just a gigantic improvement, and we expect that again for the future versions.

你看看 2.5 与 1.5 相比，这是一个巨大的改进，我们希望在未来的版本中再次实现这一点。

And so the models are becoming more capable.

因此，模型的功能变得越来越强大。

1:25:26 So you’ve got, the interesting thing about the design space in today’s world, these AI first products is you’ve got to design not for what the thing can do today, the technology can do today, but in a year’s time.

所以，当今世界设计空间的有趣之处在于，这些人工智能优先的产品是你必须设计的不是今天的东西可以做什么，技术今天可以做什么，而是一年后的事情。

So you actually have to be a very technical product person, because you’ve got to have a good intuition for and feel for, okay, that thing that I’m dreaming about now can’t be done today, but is the research track on schedule to basically intercept that in six months or a year’s time.

所以你实际上必须是一个非常技术性的产品人员，因为你必须有很好的直觉和感觉，好吧，我现在梦想的事情今天无法完成，但是按计划进行的研究轨迹基本上可以在六个月或一年的时间内拦截它。

1:25:55 So you’ve kind of got to intercept where this highly changing technology’s going, as well as the new capabilities are coming online all the time that we didn’t realize before that can allow these research to work.

因此，你必须了解这种高度变化的技术的发展方向，以及我们之前没有意识到的新功能一直在上线，这可以让这些研究发挥作用。

Or now we’ve got video generation, what do we do with that, this multimodal stuff.

或者现在我们有了视频生成功能，我们用它做什么，这种多模式的东西。

1:26:13 Is it, one question I have is it really going to be the current UI that we have today, these text box chats?

我的一个问题是，这些文本框聊天真的会成为我们今天拥有的当前用户界面吗？

Seems very unlikely once you think about these super multimodal systems.

一旦你想到这些超级多式联运系统，这似乎不太可能。

Shouldn’t it be something more like Minority Report where you are sort of vibing with it in a kind of collaborative way?

难道它不应该更像《少数派报告》那样，让你以一种协作的方式与它产生共鸣吗？

It seems very restricted today.

今天似乎受到很大限制。

I think we’ll look back on today’s interfaces and products and systems as quite archaic in maybe in just a couple of years.

我认为，也许在短短几年内，我们就会回顾今天的界面、产品和系统，认为它们已经相当过时了。

1:26:41 So I think there’s a lot of space actually for innovation to happen on the product side as well as the research side.

所以我认为在产品方面和研究方面实际上都有很大的创新空间。

And then we are offline talking about the keyboard is, the open question is how, when and how much will we move to audio as the primary way of interacting with the machines around us versus typing stuff?

然后我们离线谈论键盘时，悬而未决的问题是我们将如何、何时以及在多大程度上转向音频作为与我们周围的机器交互的主要方式而不是打字？

Yeah, I mean typing is a very low bandwidth way of doing it, even if you’re a very fast typer.

是的，我的意思是打字是一种非常低带宽的方式，即使你打字速度非常快。

1:27:00 And I think we’re going to have to start utilizing other devices, whether that’s smart glasses, audio earbuds, and eventually maybe some sorts of neural devices, where we can increase the input and the output bandwidth to something maybe a 100x of what is today.

我认为我们将不得不开始利用其他设备，无论是智能眼镜、音频耳机，还是最终可能的某种神经设备，我们可以将输入和输出带宽增加到今天的 100 倍。

I think that underappreciated art form is the interface design because I think you can not unlock the power of the intelligence of a system if you don’t have the right interface.

我认为被低估的艺术形式是界面设计，因为我认为如果没有正确的界面，就无法释放系统智能的力量。

1:27:24 The interface is really the way you unlock its power.

界面确实是您释放其力量的方式。

It’s such an interesting question of how to do that.

如何做到这一点是一个非常有趣的问题。

So how you would think getting out of the way isn’t real art form.

所以你认为如何让路并不是真正的艺术形式。

Yes.

是的。

1:27:46 It’s the sort of thing that I guess Steve Jobs always talked about, right?

我想这是史蒂夫·乔布斯经常谈论的事情，对吧？

It’s simplicity, beauty, and elegance that we want.

我们想要的是简单、美丽和优雅。

And we’re not that nobody’s there yet, in my opinion.

在我看来，我们还不是无人问津。

And that’s what I would like us to get to.

这就是我希望我们达到的目标。

1:27:58 Again, it sort of speaks to Go again as a game, the most elegant, beautiful game.

再次，它再次将围棋视为一种游戏，最优雅、最美丽的游戏。

Can you make an interface as beautiful as that?

你能做出这么漂亮的界面吗？

Actually, I think we’re going to enter an era of AI-generated interfaces that are probably personalized to you, so it fits the way that you, your aesthetic, your feel, the way that your brain works and the AI kind of generates that depending on the task.

事实上，我认为我们将进入一个人工智能生成界面的时代，这些界面可能是针对你的个性化的，所以它适合你的方式、你的审美、你的感觉、你的大脑工作方式以及人工智能根据任务生成的方式。

That feels like that’s probably the direction we’ll end up in.

感觉这可能就是我们最终的方向。

Because some people are power users and they want every single parameter on the screen, everything based perhaps me with a keyboard-based navigation and to have shortcuts for everything.

因为有些人是高级用户，他们想要屏幕上的每一个参数，一切都基于键盘导航，并且有一切的快捷方式。

1:28:25 And some people like the minimalism.

有些人喜欢极简主义。

Just hide all of that complexity.

只是隐藏所有这些复杂性。

1:28:37 Yeah, exactly.

是的，完全正确。

Yeah.

是的。

1:28:39 Well, I’m glad you have a Steve Jobs mode in you as well.

嗯，我很高兴你身上也有史蒂夫·乔布斯模式。

This is great.

这太棒了。

Einstein mode, Steve Jobs mode.

爱因斯坦模式，史蒂夫·乔布斯模式。

1:28:47 All right, let me try to trick you into answering a question.

好吧，让我试着诱骗你回答一个问题。

When will Gemini 3 come up?

双子座3什么时候出？

Is it before or after DTS-6?

是在DTS-6之前还是之后？

The world waits for both.

世界等待着两者。

1:28:56 And what does it take to go from 2.5 To 3.0?

从 2.5 到 3.0 需要什么？

Because it seems like there’s been a lot of releases of 2.5, which are already leaps in performance.

因为2.5似乎已经发布了很多版本，性能已经有了飞跃。

So what does it even mean to go to a new version?

那么升级到新版本意味着什么呢？

Is it about performance?

与性能有关吗？

Is it about a completely different flavor of an experience?

这是一种完全不同的体验吗？

Yeah, well, so the way it works with our different version numbers is we try to collect, so maybe it takes roughly six months or something to do a new kind of full run and the full productization of a new version.

是的，嗯，所以它与我们不同的版本号一起工作的方式是我们尝试收集，所以也许需要大约六个月的时间才能进行新版本的全面运行和完整产品化。

1:29:32 And during that time, lots of new interesting research iterations and ideas come up, and we sort of collect them all together that you could imagine the last six months worth of interesting ideas on the architecture front, maybe it’s on the data front, it’s like many different possible things.

在那段时间里，出现了许多新的有趣的研究迭代和想法，我们将它们全部收集在一起，你可以想象过去六个月在架构方面的有趣想法，也许是在数据方面，就像许多不同的可能的事情。

And we package that all up, test which ones are likely to be useful for the next iteration, and then bundle that all together.

我们将它们全部打包，测试哪些可能对下一次迭代有用，然后将它们捆绑在一起。

And then we start the new giant hero training run.

然后我们开始新的巨人英雄训练。

And then of course that gets monitored.

当然，这会受到监控。

1:30:00 …

……

run, right?

跑吧？

And then of course that gets monitored and then at the end of the pre-training, then there’s all the post-training, there’s many different ways of doing that, different ways of patching it.

当然，这会受到监控，然后在预训练结束时，然后是所有的后训练，有很多不同的方法可以做到这一点，有不同的修补方法。

So there’s a whole experimental phase there which you can also get a lot of gains out.

因此，这里有一个完整的实验阶段，您也可以从中获得很多收益。

And that’s where you see the version numbers usually referring to the base model, the pre-trained model, and then the interim versions of 2.5 and the different sizes and the different little additions.

在那里你可以看到版本号，通常指的是基础模型、预训练模型，然后是 2.5 的临时版本以及不同的大小和不同的小添加。

They’re often patches or post-training ideas that can be done afterwards off the same basic architecture.

它们通常是补丁或训练后的想法，可以在相同的基本架构上完成。

And then of course on top of that, we also have different sizes, Pro and Flash and Flashlight that are often distilled from the biggest ones, the Flash model from the Pro model.

当然，除此之外，我们还有不同的尺寸，Pro、Flash 和 Flashlight，它们通常是从最大的型号中提取出来的，Flash 型号是从 Pro 型号中提取出来的。

And that means we have a range of different choices.

这意味着我们有一系列不同的选择。

If you’re the developer, do you want to prioritize performance or speed and cost?

如果您是开发人员，您想优先考虑性能还是速度和成本？

1:30:51 And we like to think of this Pareto frontier of on the one hand, the Y-axis is like performance, and then the X- axis is cost or latency and speed basically.

我们喜欢考虑这个帕累托边界，一方面，Y 轴就像性能，然后 X 轴基本上是成本或延迟和速度。

And we have models that completely define the frontier.

我们拥有完全定义边界的模型。

So whatever your trade-off is that you want as an individual user or as a developer, you should find one of our models satisfies that constraint.

因此，无论您作为个人用户还是开发人员想要进行权衡，您都应该发现我们的模型之一满足该约束。

So behind the version changes, there is a big run and then there’s just an insane complexity of productization.

因此，在版本更改的背后，有一个大的运行，然后就是产品化的疯狂复杂性。

1:31:17 Then there’s the distillation of the different sizes along that Pareto front.

然后是沿着帕累托前沿的不同尺寸的蒸馏。

And then as with each step you take, you realize there might be a cool product.

然后，随着您迈出的每一步，您都会意识到可能会有一个很酷的产品。

There’s side quests.

有支线任务。

Yes, exactly.

是的，完全正确。

And then you also don’t want to take too many side quests because then you have a million versions and a million products.

然后你也不想接受太多的支线任务，因为这样你就有了一百万个版本和一百万个产品。

Yes, precisely.

是的，正是如此。

It’s very unclear, but you also get super excited because it’s super cool.

虽然还不清楚，但你也会感到非常兴奋，因为它非常酷。

1:31:46 How does even look at VLs?

如何看待 VL？

Very cool.

非常酷。

How does it fit into the bigger Thing?

它如何适应更大的事物？

Yes, exactly.

是的，完全正确。

1:31:55 Exactly.

确切地。

And then you’re constantly this process of converging upstream, we call it ideas from the product surfaces or from the post-training and even further downstream and that, you upstream that into the core model training for the next run.

然后你不断地向上游汇聚的过程，我们称之为来自产品表面或训练后甚至更下游的想法，你将其上游到下一次运行的核心模型训练中。

So then the main model, the main Gemini track becomes more and more general and eventually, AGI.

所以主要模型，主要的双子座轨道变得越来越通用，最终，AGI。

One hero run.

一位英雄奔跑。

Yes, exactly.

是的，完全正确。

1:32:21 A few hero runs later.

一些英雄随后逃跑。

Yeah.

是的。

1:32:23 So sometimes when you release these new versions or every version, really, are benchmarks productive or counterproductive for showing the performance of a model?

因此，有时当您发布这些新版本或每个版本时，基准测试对于显示模型的性能真的是有效还是适得其反？

You need them, but it’s important that you don’t overfit to them.

您需要它们，但重要的是不要过度适应它们。

1:32:36 So they shouldn’t be the be all and end all.

所以它们不应该是一切。

So there’s LMArena, or it used to be called LEMSYS, that’s one of them that turned out organically to be one of the main ways people like to test these systems, at least the chatbots.

所以就有了 LMArena，或者它曾经被称为 LEMSYS，它是其中之一，最终成为人们喜欢测试这些系统（至少是聊天机器人）的主要方式之一。

Obviously there’s loads of academic benchmarks that test mathematics and coding ability, general language ability, science ability and so on.

显然，有大量的学术基准测试数学和编码能力、通用语言能力、科学能力等。

And then we have our own internal benchmarks that we care about.

然后我们有自己关心的内部基准。

1:33:04 It’s a multi objective optimization problem.

这是一个多目标优化问题。

You don’t want to be good at just one thing.

你不想只擅长一件事。

We’re trying to build general systems that are good across the board, and you try and make no-regret improvements.

我们正在尝试构建全面良好的通用系统，并且您尝试做出无悔的改进。

So where you improve in coding, but it doesn’t reduce your performance in other areas.

因此，您在编码方面有所改进，但不会降低您在其他领域的表现。

So that’s the hard part because of course you could put more coding data in or you could put more, I don’t know, gaming data in, but then does it make worse your language system or your translation systems and other things that you care about?

所以这是困难的部分，因为你当然可以放入更多的编码数据，或者你可以放入更多的，我不知道，游戏数据，但是这会让你的语言系统或翻译系统以及你关心的其他东西变得更糟吗？

So you’ve got to continually monitor this increasingly larger and larger suite of benchmarks.

因此，您必须持续监控这套越来越大的基准测试。

And also when you stick them into products, these models, you also care about the direct usage and the direct stats and the signals that you’re getting from the end users, whether they’re coders or the average person using the chat interfaces.

而且，当您将它们应用到产品、这些模型中时，您还会关心直接使用情况、直接统计数据以及从最终用户那里获得的信号，无论他们是程序员还是使用聊天界面的普通人。

Because ultimately, you want to measure the usefulness, but it’s so hard to convert that into a number.

因为最终，您想要衡量有用性，但很难将其转换为数字。

1:34:00 It’s really vibe based benchmarks across a large number of users.

这确实是针对大量用户的基于氛围的基准测试。

And it’s hard to know and it would be just terrifying to me, you know have a much smarter model, but it’s just something vibe based.

这很难知道，这对我来说会很可怕，你知道有一个更聪明的模型，但这只是基于氛围的东西。

It’s not quite working.

这不太有效。

That’s such a scary and everything you just said.

你刚才所说的一切真是太可怕了。

It has to be smart and useful across so many domains.

它必须在如此多的领域中智能且有用。

So you get super excited all of a sudden solving programming problems you’ve never been able to solve before, but now it’s crappier poetry or something and it’s just, I don’t know, that’s a stressful.

所以你会突然变得超级兴奋，解决你以前从未解决过的编程问题，但现在它是更蹩脚的诗歌或其他东西，只是，我不知道，这是一种压力。

That’s so difficult-

这太难了——

To balance.

为了平衡。

To balance and because you can’t really trust the benchmarks, you really have to trust the end users.

为了平衡，因为你不能真正信任基准，所以你真的必须信任最终用户。

Yeah.

是的。

1:34:48 And then other things that are even more esoteric come into play, like the style of the persona of the system, is it verbose?

然后其他更深奥的事情开始发挥作用，比如系统角色的风格，是否冗长？

Is it succinct?

是否简洁？

Is it humorous?

是不是很幽默？

And different people like different things.

而且不同的人喜欢不同的东西。

So it’s very interesting.

所以这很有趣。

It’s almost like cutting edge part of psychology research or personality research.

这几乎就像心理学研究或人格研究的前沿部分。

I used to do that in my PhD, like five factor personality, what do we actually want our systems to be like?

我曾经在我的博士学位中这样做过，就像五因素人格一样，我们实际上希望我们的系统是什么样的？

And different people will like different things as well.

而且不同的人也会喜欢不同的东西。

So these are all just new problems in product space that I don’t think I’ve ever really been tackled before, but we’re going to rapidly have to deal with now.

因此，这些都是产品领域的新问题，我认为以前从未真正解决过这些问题，但我们现在必须迅速解决。

I think it’s a super fascinating space, developing the character of the thing and in so doing, it puts a mirror to ourselves, what are the kind of things that we like?

我认为这是一个超级迷人的空间，发展事物的特征，这样做，它给我们自己树立了一面镜子，我们喜欢什么样的事物？

1:35:27 Because prompt engineering allows you to control a lot of those elements, but can the product make it easier for you to control the different flavors of those experiences, the different characters that you interact with?

因为即时工程允许您控制其中的许多元素，但是该产品能否使您更轻松地控制这些体验的不同风格以及与您互动的不同角色？

Yeah, exactly.

是的，完全正确。

So what’s the probability of Google DeepMind winning?

那么 Google DeepMind 获胜的概率有多大呢？

Well, I see it as winning.

嗯，我认为这是胜利。

1:35:56 I think winning is the wrong way to look at it given how important and consequential what it is we’re building.

我认为，考虑到我们正在构建的东西的重要性和影响力，以获胜的方式来看待它是错误的。

So funny enough, I try not to view it like a game or competition even though that’s a lot of my mindset.

很有趣的是，我尽量不把它当作一场游戏或比赛，尽管这是我的很多心态。

It’s about in my view, all of us or those of us at the leading edge or have a responsibility to steward this unbelievable technology that could be used for incredible good but also has risks, steward it safely into the world for the benefit of humanity.

在我看来，我们所有人或我们中处于领先地位的人或有责任管理这项令人难以置信的技术，它可以用于令人难以置信的好处，但也存在风险，为了人类的利益，安全地管理它进入世界。

That’s always what I’ve dreamed about and what we’ve always tried to do.

这一直是我的梦想，也是我们一直努力去做的事情。

And I hope that’s what eventually the community, maybe the international community will rally around when it becomes obvious that as we get closer and closer to AGI, that’s what’s needed.

我希望这就是最终社区，也许国际社会会团结起来，当我们越来越接近通用人工智能时，这就是我们所需要的。

I agree with you.

我同意你的看法。

1:36:43 I think that’s beautifully put.

我认为这说得很漂亮。

You’ve said that you talk to and are on good terms with the leads of some of these labs.

您说过您与其中一些实验室的领导进行了交谈并保持着良好的关系。

As the competition heats up, how hard is it to maintain those relationships?

随着竞争的加剧，维持这些关系有多难？

It’s been okay so far.

到目前为止还好。

1:36:58 I try to pride myself in being collaborative.

我努力为自己的协作能力感到自豪。

I’m a collaborative person.

我是一个善于合作的人。

Research is a collaborative endeavor.

研究是一项协作努力。

Science is a collaborative endeavor.

科学是一项协作努力。

It’s all good for humanity in the end if you cure terrible diseases and you come up with an incredible cure, this is net win for humanity.

如果你治愈了可怕的疾病并且想出了令人难以置信的治疗方法，那么最终对人类来说都是有好处的，这对人类来说是净胜利。

And the same with energy, all of the things that I’m interested in helping solve with AI.

能源也是如此，我有兴趣通过人工智能帮助解决所有问题。

So I just want that technology to exist in the world and be used for the right things and the benefits of that, the productivity benefits of that being shared for the benefit of everyone.

所以我只是希望这项技术能够存在于世界上，并被用于正确的事情，并从中受益，其生产力的好处可以为每个人的利益而共享。

So I try to maintain good relations with all the leading lab people.

因此，我努力与实验室的所有领导人员保持良好的关系。

They’re very interesting characters, many of them as you might expect.

他们都是非常有趣的角色，其中很多都如你所料。

Yep.

是的。

But yeah, I’m on good terms I hope with pretty much all of them.

但是，是的，我希望我和他们所有人的关系都很好。

1:37:41 And I think that’s going to be important when things get even more serious than they are now, that there are those communication channels and that’s what will facilitate cooperation or collaboration if that’s what is required, especially on things like safety.

我认为，当事情变得比现在更加严重时，这些沟通渠道将很重要，如果需要的话，这将促进合作或协作，尤其是在安全等方面。

Yeah, I hope there’s some collaboration on stuff that’s less high stakes and in so doing, serves as a mechanism for maintaining friendships and relationships.

是的，我希望在风险较低的事情上进行一些合作，这样可以作为维持友谊和关系的机制。

1:38:00 So for example, I think the internet would love it if you and Elon somehow collaborate on creating a video game, that kind of thing.

举例来说，我认为如果你和埃隆以某种方式合作创建视频游戏之类的事情，互联网会喜欢它。

I think that enables camaraderie and good terms.

我认为这可以带来友情和良好的关系。

And also you two are legit gamers, so it’s just fun to to create some-

而且你们两个都是合法的游戏玩家，所以创造一些东西很有趣-

Yeah, that would be awesome.

是的，那太棒了。

1:38:22 And we’ve talked about that in the past and it may be a cool thing that we can do.

我们过去已经讨论过这一点，我们可以做的这可能是一件很酷的事情。

And I agree with you, it’d be nice to have side projects in a way where one can just lean into the collaboration aspect of it and it’s a a win-win for both sides and it builds up that collaborative muscle.

我同意你的观点，如果能有一些副业项目，让人们可以专注于它的协作方面，这对双方来说都是双赢的，并且可以增强协作能力。

I see the scientific endeavor as that side project for humanity and I think Google DeepMind has been really pushing that.

我将科学事业视为人类的副业项目，并且我认为 Google DeepMind 一直在真正推动这一目标。

1:38:44 I would love to see other labs do more scientific stuff and then collaborate because it just seems like it’s easier to collaborate on the big scientific questions.

我很高兴看到其他实验室做更多的科学工作，然后进行合作，因为在重大科学问题上合作似乎更容易。

I agree and I would love to see a lot of people, all of the other labs talk about science, but I think we’re really the only ones using it for science and doing that.

我同意，我很乐意看到很多人，所有其他实验室都在谈论科学，但我认为我们确实是唯一将其用于科学并这样做的人。

1:39:01 And that’s why projects like AlphaFold are so important to me.

这就是为什么像 AlphaFold 这样的项目对我来说如此重要。

And I think to our mission is to show how AI can be clearly used in a very concrete way for the benefit of humanity.

我认为我们的使命是展示如何以非常具体的方式明确使用人工智能来造福人类。

And also, we spun out companies like Isomorphic off the back of Alpha Fold to do drug discovery and it’s going really well and you can think of build additional AlphaFold type systems to go into chemistry space to help accelerate drug design.

而且，我们在 Alpha Fold 的基础上剥离了像 Isomorphic 这样的公司来进行药物发现，进展非常顺利，你可以考虑构建额外的 AlphaFold 类型系统进入化学领域，以帮助加速药物设计。

And the examples I think we need to show and society needs to understand are where AI can bring these huge benefits.

我认为我们需要展示、社会需要理解的例子是人工智能可以带来这些巨大好处的地方。

Well, from the bottom of my heart, thank you for pushing the scientific efforts forward with rigor, with fun, with humility, all of it.

好吧，我从心底里感谢你们以严谨、有趣、谦虚的态度推动科学事业的发展。

1:39:42 I just love to see and still talking about P equals NP, it’s just incredible.

我只是喜欢看到并且仍然谈论 P 等于 NP，这真是令人难以置信。

So I love it.

所以我喜欢它。

There’s been seemingly a war for talent.

似乎存在着一场人才争夺战。

Some of it is meme, I don’t know.

有些是模因，我不知道。

What do you think about Meta buying up talent with huge salaries and the heating up of this battle for talent?

对于Meta高薪收购人才，以及这场人才争夺战的升温，您有何看法？

I should say that I think a lot of people see DeepMind as a really great place to do cutting-edge work for the reasons that you’ve outlined.

我应该说，我认为很多人都将 DeepMind 视为进行前沿工作的好地方，原因正如您所概述的。

There’s this vibrant scientific culture.

这里有充满活力的科学文化。

Yeah.

是的。

1:40:21 Well look, of course there’s a strategy that Meta is taking right now.

好吧，Meta 现在当然正在采取一个策略。

I think that from my perspective at least, I think the people that are real believers in the mission of AGI and what it can do and understand the real consequences, both good and bad from that and what that responsibility entails, I think they’re mostly doing it to be like myself, to be on the frontier of that research so they can help influence the way that goes and steward that technology safely into the world.

我认为至少从我的角度来看，我认为那些真正相信 AGI 使命和它能做什么、了解真正后果的人，无论好坏，以及责任所带来的后果，我认为他们这样做主要是为了像我一样，站在这项研究的前沿，这样他们就可以帮助影响发展方向，并管理这项技术安全地进入世界。

And Meta right now are not at the frontier.

而 Meta 目前还没有处于前沿。

Maybe they’ll manage to get back on there and it’s probably rational what they’re doing from their perspective because they’re behind and they need to do something.

也许他们会设法回到那里，从他们的角度来看，他们所做的事情可能是合理的，因为他们落后了，他们需要做点什么。

But I think there’s more important things than just money.

但我认为还有比钱更重要的东西。

Of course one has to pay people their market rates and all of these things and that continues to go up.

当然，人们必须向人们支付市场价格以及所有这些费用，而且这种情况还在继续上升。

And I was expecting this because more and more people are finally realizing, leaders of companies, what I’ve always known for 30 plus years now, which is that AGI is the most important technology probably that’s ever going to be invented.

我对此有所期待，因为越来越多的人、公司领导者终于意识到，我 30 多年来一直知道的一点是，AGI 可能是有史以来最重要的技术。

So in some senses, it’s rational to be doing that.

所以从某种意义上说，这样做是合理的。

But I also think there’s a much bigger question.

但我也认为还有一个更大的问题。

People in AI these days are very well paid.

现在人工智能领域的人薪水很高。

1:41:32 I remember when we were starting out back in 2010, I didn’t even pay myself a couple of years because it wasn’t enough money.

我记得 2010 年我们刚开始创业时，有几年我什至没有给自己发工资，因为钱不够。

We couldn’t raise any money, and these days, interns are being paid the amount that we raised as our first entire seed round.

我们无法筹集到任何资金，而如今，实习生的工资是我们在第一轮种子轮融资中筹集到的金额。

So it’s pretty funny.

所以这很有趣。

And I remember the days where I used to have to work for free and almost pay my own way to do an internship.

我还记得以前我必须免费工作，几乎要自己掏钱去实习的日子。

Right now, it’s all the other around, but that’s just how it is.

现在，情况完全不同，但事实就是如此。

It’s the new world.

这是新世界。

But I think that we’ve been discussing what happens post- AGI and energy systems are solved and so on, what is even money going to mean?

但我认为我们一直在讨论通用人工智能和能源系统解决后会发生什么等等，甚至金钱意味着什么？

So I think in the economy and we’re going to have much bigger issues to work through and how does the economy function in that world and companies?

因此，我认为在经济领域，我们将有更大的问题需要解决，经济在这个世界和公司中如何运作？

So I think it’s a little bit of a side issue about salaries and things like that today.

所以我认为今天关于工资之类的事情有点次要问题。

Yeah, when you’re facing such gigantic consequences and gigantic, fascinating scientific questions-

是的，当你面临如此巨大的后果和巨大的、令人着迷的科学问题时——

Which may be only a few years away.

这可能只有几年的时间了。

So the practicals, the pragmatic sense, if we zoom in on jobs, we can look at programmers because it seems like AI systems are currently doing incredibly well at programming and increasingly so.

因此，从实际情况来看，务实的感觉，如果我们放大工作，我们可以关注程序员，因为人工智能系统目前在编程方面似乎做得非常好，而且越来越好。

1:42:27 So A lot of people that program for a living, love programming are worried they will lose their jobs.

所以很多以编程为生、热爱编程的人担心他们会失去工作。

How worried should they be do you think, and what’s the right way to adjust to the new reality and ensure that you survive and thrive as a human in the programming world?

您认为他们应该有多担心？适应新现实并确保您作为一个人在编程世界中生存和发展的正确方法是什么？

Well, it’s interesting that programming, and it’s again counterintuitive to what we thought years ago, maybe that some of the skills that we think of as harder skills are turned out maybe to be the easier ones for various reasons.

嗯，编程很有趣，这又与我们多年前的想法相反，也许我们认为较难的一些技能由于各种原因可能变得更容易。

1:42:58 But coding and maths, because you can create a lot of synthetic data and verify if that data’s correct.

但是编码和数学，因为您可以创建大量合成数据并验证这些数据是否正确。

So because of that nature of that, it’s easier to make things like synthetic data to train from.

因此，由于其本质，更容易制作诸如合成数据之类的数据来进行训练。

It’s also an area of course we’re all interested in because as programmers to help us and get faster at it and more productive.

当然，这也是我们都感兴趣的一个领域，因为作为程序员可以帮助我们更快地完成工作并提高工作效率。

1:43:27 So I think for the next era, like the next five, 10 years, I think what we’re going to find is people who embrace these technologies become almost at one with them, whether that’s in the creative industries or the technical industries will become superhumanly productive, I think.

因此，我认为在下一个时代，比如未来五年、十年，我认为我们将发现拥抱这些技术的人们几乎与它们融为一体，无论是在创意产业还是在技术产业，我认为都将变得超人的生产力。

So the great programmers will be even better, but there’ll be even 10X even what they are today.

所以伟大的程序员会更好，但甚至会比现在高出 10 倍。

And because there, you’ll be able to use their skills to utilize the tools to the maximum, exploit them to the maximum.

因为在那里，你将能够利用他们的技能来最大限度地利用这些工具，最大限度地利用它们。

And so I think that’s what we’re going to see in the next domain.

所以我认为这就是我们将在下一个领域看到的。

So that’s going to cause quite a lot of change.

所以这会带来很大的变化。

And so that’s coming.

所以这即将到来。

A lot of people benefit from that.

很多人都从中受益。

1:44:05 So I think one example of that is if coding becomes easier, it becomes available to many more creatives to do more.

因此，我认为一个例子是，如果编码变得更加容易，那么更多的创意人员就可以做更多的事情。

But I think the top programmers will still have huge advantages as terms of specifying, going back to specifying what the architecture should be.

但我认为顶级程序员在指定方面仍然具有巨大的优势，回到指定架构应该是什么。

The question should be how to guide these coding assistants in a way that’s useful and check whether the code they produce is good.

问题应该是如何以有用的方式指导这些编码助手并检查他们生成的代码是否良好。

So I think there’s plenty of headroom there for the foreseeable next few years.

所以我认为在可预见的未来几年里还有很大的发展空间。

So I think there’s several interesting things there.

所以我认为那里有一些有趣的事情。

1:44:36 One is there’s a lot of imperative to just get better and better consistently of using these tools so they’re riding the wave of the improving models versus competing against them.

一是必须不断地更好地使用这些工具，这样他们才能顺应模型改进的浪潮，而不是与模型竞争。

But sadly, but that’s the nature of life on earth, there could be a huge amount of value to certain kinds of programming at the cutting edge and less value to other kinds.

但可悲的是，这就是地球生命的本质，某些前沿的编程可能具有巨大的价值，而其他类型的价值则较小。

For example, it could be front-end web design might be more amenable to, as you’ve mentioned, to generation by AI systems and maybe for example, game engine design or something like this or back-end design or guiding systems in high-performance situations, high-performance programming type of design decisions, that might be extremely valuable.

例如，正如您所提到的，前端网页设计可能更适合由人工智能系统生成，例如游戏引擎设计或类似的东西，或者高性能情况下的后端设计或指导系统，高性能编程类型的设计决策，这可能非常有价值。

But it will shift where the humans are needed most and that’s scary for people to address.

但它会改变最需要人类的地方，而这对人们来说是可怕的。

Yeah, I think that’s right.

是的，我认为这是对的。

1:45:37 Anytime where there’s a lot of disruption and change, and we’ve had this, it’s not just this time.

任何时候都会有很多干扰和变化，我们也遇到过这种情况，但不仅仅是这一次。

We’ve had this in many times in human history with the internet, mobile, but before that obviously, the Industrial Revolution and it’s going to be one of those eras where there will be a lot of change.

在人类历史上，我们已经通过互联网、移动设备多次经历过这种情况，但显然在此之前是工业革命，这将是发生很多变化的时代之一。

I think there’ll be new jobs we can’t even imagine today, just like the internet created.

我认为将会出现我们今天无法想象的新工作，就像互联网创造的那样。

And then those people with the right skill sets to ride that wave will become incredibly valuable, those skills.

然后，那些拥有适当技能来驾驭这股浪潮的人将变得非常有价值，这些技能。

But maybe people will have to relearn or adapt a bit, their current skills.

但也许人们将不得不重新学习或适应他们当前的技能。

And the thing that’s going to be harder to deal with this time around is that I think what we’re going to see is something like probably 10 times the impact the Industrial Revolution had, but 10 times faster as well.

这次更难处理的事情是，我认为我们将看到的影响可能是工业革命的 10 倍，但速度也快 10 倍。

So instead of a 100 years, it takes 10 years and so that’s going to make, it’s like a 100X, the impact and the speed combined.

因此，它需要 10 年，而不是 100 年，所以这就像 100 倍，影响和速度相结合。

1:46:31 So I think going to make it more difficult for society to deal with and there’s a lot to think through and I think we need to be discussing that right now.

所以我认为这会让社会更加难以应对，有很多事情需要思考，我认为我们现在需要讨论这个问题。

And I encourage top economists in the world and philosophers to start thinking about how is society going to be affected by this and what should we do?

我鼓励世界顶级经济学家和哲学家开始思考社会将如何受到影响以及我们应该做什么？

Including things like universal basic provision or something like that where a lot of the increased productivity gets shared out and distributed to society and maybe in the form of services and other things where if you want more than that, you still go and get some incredibly rare skills and things like that and make yourself unique.

包括诸如普遍的基本供给之类的东西，其中许多提高的生产力被分享并分配给社会，也许以服务和其他东西的形式，如果你想要更多，你仍然可以去获得一些极其罕见的技能和类似的东西，让自己独一无二。

But there’s a basic provision that is provided.

但提供了一个基本规定。

And if you think of government as a technology, there’s also interesting questions, not just in the economics, but just politics.

如果你将政府视为一种技术，那么就会存在一些有趣的问题，不仅在经济方面，而且在政治方面。

1:47:19 How do you design a system that’s responding to the rapidly changing times such that you can represent the different pain that people feel from the different groups and how do you reallocate resources in a way that addresses that pain and represents the hope and the pain and the fears of different people in a way that doesn’t lead to division?

如何设计一个能够应对快速变化的时代的系统，以便能够代表不同群体的人们所感受到的不同痛苦，以及如何以一种不会导致分裂的方式来解决这种痛苦并代表不同人的希望、痛苦和恐惧的方式重新分配资源？

Because politicians are often really good at fueling the division and using that to get elected, defining the other and then saying that’s bad.

因为政客们往往非常擅长助长分歧，并利用分歧来当选，定义对方，然后说那很糟糕。

And based on that, I think that’s often counterproductive to leveraging a rapidly changing technology to help the world flourish.

基于此，我认为这对于利用快速变化的技术来帮助世界繁荣往往会适得其反。

So we almost need to improve our political systems as well rapidly, if you think of them as a technology.

因此，如果你将政治体系视为一项技术，我们几乎也需要迅速改善它们。

Definitely.

确实。

1:48:19 And I think we’ll need new governance structures, institutions probably to help with this transition.

我认为我们需要新的治理结构和机构来帮助实现这一转变。

So I think political philosophy and political science is going to be key to that.

所以我认为政治哲学和政治学将是实现这一目标的关键。

But I think the number one thing, first of all is to create more abundance of resources.

但我认为第一件事，首先是创造更丰富的资源。

So that’s the number one thing.

所以这是第一件事。

Increase productivity, get more resources, maybe eventually get out of the zero-sum situation.

提高生产力，获得更多资源，也许最终摆脱零和局面。

Then the second question is how to use those resources and distribute those resources.

那么第二个问题就是如何使用这些资源并分配这些资源。

But yeah, you can’t do that without having that abundance first.

但是，是的，如果没有足够的资源，你就无法做到这一点。

You mentioned to me the book, The Maniac by Benjamin Labatut, a book on first of all about you.

你向我提到了本杰明·拉巴图（Benjamin Labatut）的《疯子》（The Maniac）一书，这本书首先是关于你的。

1:48:54 There’s a bio about you.

有一份关于你的简历。

Strange, yeah.

奇怪，是的。

Yes, sure.

是的，当然。

1:49:06 It’s unclear how much is fiction, how much is reality.

分不清有多少是虚构的，有多少是现实。

But I think the central figure that is John von Neumann, I would say it’s a haunting and beautiful exploration of madness and genius and let’s say the double-edged sword of discovery.

但我认为中心人物是约翰·冯·诺依曼，我想说这是对疯狂和天才的一次令人难忘的美丽探索，可以说是发现的双刃剑。

And for people who don’t know, John von Neumann is a legendary mind.

对于那些不知道的人来说，约翰·冯·诺依曼是一位传奇人物。

He contributed to quantum mechanics.

他对量子力学做出了贡献。

He was on the Manhattan Project.

他参与了曼哈顿计划。

He is widely considered to be the father of or pioneer the modern computer and AI and so on.

他被广泛认为是现代计算机和人工智能等的之父或先驱。

So many people say he’s one of the smartest humans ever, which is fascinating.

很多人说他是有史以来最聪明的人之一，这很有趣。

1:49:45 And what’s also fascinating is he’s a person who saw nuclear science and physics become the atomic bomb, so you got to see ideas become a thing that has a huge amount of impact on the world.

同样令人着迷的是，他见证了核科学和物理学成为原子弹，所以你必须看到想法成为对世界产生巨大影响的事物。

He also foresaw the same thing for computing, and that’s a little bit again, beautiful and haunting aspect of the book.

他还预见到计算也会发生同样的事情，这又是这本书美丽而令人难以忘怀的方面。

Then taking a leap forward and looking at this, at least it all AlphaZero, AlphaGo AlphaZero big moment that maybe John von Neumann’s thinking was brought to reality.

然后向前跃进看看这一点，至少这一切都是AlphaZero、AlphaGo AlphaZero的伟大时刻，也许约翰·冯·诺依曼的想法被变成了现实。

So I guess the question is what do you think if you got to hang out with John von Neumann now, what would he say about what’s going on?

所以我想问题是，如果你现在和约翰·冯·诺依曼一起出去玩，你怎么想，他会对正在发生的事情说些什么？

Well, that would be an amazing experience.

嗯，那将是一次奇妙的经历。

1:50:35 He’s a fantastic mind.

他的头脑非常出色。

And I also love the way he spent a lot of his time at Princeton at the Institute of Advanced Studies, a very special place for thinking.

我也喜欢他在普林斯顿大学高级研究院度过的大部分时间，这是一个非常特别的思考场所。

And it’s amazing how much of a polymath he was and the spread of things he helped invent, including of course the Von Neumann architecture that all the modern computers are based on.

令人惊讶的是，他是一位博学多才的人，他帮助发明的东西得以传播，当然包括所有现代计算机所基于的冯·诺依曼架构。

And he had amazing foresight.

他有惊人的远见。

I think he would’ve loved where we are today, and I think he would’ve really enjoyed AlphaGo being, he did game theory.

我认为他会喜欢我们今天的处境，我认为他会非常喜欢 AlphaGo，他研究博弈论。

I think he foresaw a lot of what would happen with learning machines, systems that are grown, I think he called it rather than programmed.

我认为他预见到了学习机器和成长的系统会发生的很多事情，我认为他称之为而不是编程。

I’m not sure how even maybe he wouldn’t even be that surprised.

我不确定他怎么可能不会那么惊讶。

There’s the fruition of what I think he already foresaw in the 1950s.

我认为他在 20 世纪 50 年代就已经预见到了这一点。

I wonder what advice he would give.

我想知道他会给出什么建议。

1:51:25 He got to see the building of the atomic bomb with the Manhattan Project.

他通过曼哈顿计划亲眼目睹了原子弹的制造。

I’m sure there’s interesting stuff that maybe is not talked about enough, maybe some bureaucratic aspect, maybe the influence of politicians, maybe not enough of picking up the phone and talking to people that are called enemies by the said politicians.

我确信有一些有趣的事情可能没有被充分讨论，也许是一些官僚方面，也许是政客的影响，也许是没有足够的拿起电话与那些被政客称为敌人的人交谈。

There might be some deep wisdom that we just may have lost from that time actually.

实际上，我们可能从那时起就失去了一些深刻的智慧。

Yeah, I’m sure there is.

是的，我确信有。

1:51:48 I read a lot of books for that time as well, Chronicle Time and some brilliant people involved.

我也读了很多关于那个时期的书，《编年史》和一些参与其中的杰出人物。

But I agree with you.

但我同意你的观点。

I think maybe there needs to be more dialogue and understanding.

我认为也许需要更多的对话和理解。

I hope we can learn from those times.

我希望我们能从那些时代中吸取教训。

I think the difference here is that the AI has so many, it’s a multi-use technology.

我认为这里的区别在于人工智能有很多，它是一种多用途技术。

Obviously we’re trying to do things like solve all diseases, help with energy and scarcity, these incredible things.

显然，我们正在努力做一些事情，比如解决所有疾病、帮助解决能源和稀缺问题，这些令人难以置信的事情。

This is why all of us and myself, I started on this journey 30 plus years ago.

这就是为什么我们所有人和我自己在 30 多年前开始了这段旅程。

But of course there are risks too.

但当然也存在风险。

And probably Von Neumann, my guess is he foresaw both.

我猜冯·诺依曼可能预见到了这两点。

And I think he said, I think it’s to his wife, that computers would be even more impactful in the world.

我想他是对他的妻子说的，计算机会对世界产生更大的影响。

And as we just discussed, I think that’s right.

正如我们刚才讨论的，我认为这是对的。

I think it’s going to be 10 times at least of the Industrial Revolution.

我认为这至少是工业革命时期的 10 倍。

So I think he’s right.

所以我认为他是对的。

So I think he would’ve been, I imagine, fascinated by where we are now.

所以我想他会对我们现在的处境着迷。

And I think one of the, maybe you can correct me, but one of the takeaways from the book is that reason, as said in the book, Mad Dreams of Reason, it’s not enough for guiding humanity as we build these super powerful technology.

我认为其中之一，也许你可以纠正我，但这本书的要点之一是，正如《理性的疯狂梦想》一书中所说，当我们构建这些超级强大的技术时，这不足以指导人类。

1:52:53 That there’s something else.

那还有别的东西。

There’s also a religious component, whatever God, whatever religion gives, it pulls at something in the human spirit that raw cold reason doesn’t give us.

还有一个宗教成分，无论上帝、无论宗教给予什么，它都会在人类精神中吸引一些原始冷酷的理性无法给予我们的东西。

And I agree with that.

我同意这一点。

1:53:22 I think we need to approach it with whatever you want to call it, a spiritual dimension or humanist dimension.

我认为我们需要用任何你想称之为的东西来处理它，精神维度或人文主义维度。

Doesn’t have to be to do with religion, but this idea of a soul, what makes us human, this spark that we have, perhaps it’s to do with consciousness when we finally understand that, I think that has to be at the heart of the endeavor.

不一定与宗教有关，但灵魂的概念，是什么让我们成为人类，我们拥有的这种火花，也许与意识有关，当我们最终理解这一点时，我认为这必须是努力的核心。

And technology, I’ve always seen technology as the enabler, the tools that enable us to flourish and to understand more about the world.

至于技术，我一直将技术视为推动者，是使我们能够蓬勃发展并更多地了解世界的工具。

And I’m with Feynman on this, and he used to always talk about science and art being companions.

在这一点上我同意费曼，他过去总是谈论科学和艺术是伴侣。

You can understand it from both sides, the beauty of a flower, how beautiful it is, and also understand why the colors of the flower evolve like that.

你可以从两个方面来理解，一朵花的美丽，它是多么美丽，也可以理解为什么花的颜色会这样演变。

That just makes it more beautiful, just the intrinsic beauty of the flower.

这只会让它变得更加美丽，这就是花的内在美。

1:54:10 I’ve always seen it like that.

我一直都是这样看的。

And maybe in the Renaissance times, the great discoverers then, people like Da Vinci, I don’t think he saw any difference between science and art and perhaps religion.

也许在文艺复兴时期，伟大的发现者，像达芬奇这样的人，我认为他没有看到科学和艺术，也许还有宗教之间有任何区别。

Everything was, it’s just part of being human and being inspired about the world around us.

一切都是，这只是人类的一部分，也是我们受到周围世界启发的一部分。

And that’s the philosophy I tried to take.

这就是我试图遵循的哲学。

And one of my favorite philosophers is Spinoza.

我最喜欢的哲学家之一是斯宾诺莎。

And I think he combined that all very well, this idea of trying to understand the universe and understanding our place in it.

我认为他很好地将这一切结合起来，试图理解宇宙并理解我们在其中的地位。

And that was his way of understanding religion.

这就是他理解宗教的方式。

And I think that’s quite beautiful.

我认为这非常漂亮。

And for me, all of these things are related, interrelated, the technology and what it means to be human.

对我来说，所有这些事情都是相关的、相互关联的、技术以及人类的意义。

1:54:53 And I think it’s very important though that we remember that as when we’re immersed in the technology and the research, I think a lot of researchers that I see in our field are a little bit too narrow and only understand the technology.

我认为非常重要的是，我们要记住，当我们沉浸在技术和研究中时，我认为我在我们的领域看到的很多研究人员都有点太狭隘了，只了解技术。

And I think also that’s why it’s important for this to be debated by society at large.

我认为这也是为什么整个社会对此进行辩论很重要的原因。

I’m very supportive of things like the AI summits that will happen and governments understanding it.

我非常支持即将举行的人工智能峰会以及政府的理解等事情。

And I think that’s one good thing about the chatbot era and the product era of AI is that everyday person can actually feel and interact with cutting edge AI and feel it for themselves.

我认为聊天机器人时代和人工智能产品时代的一件好事是，每个人都可以真正感受到尖端人工智能并与之互动，并亲自感受它。

Yeah, because they force the technologists to have the human conversation.

是的，因为他们迫使技术人员进行人际对话。

1:55:30 Yeah, for sure.

是的，当然。

Yeah.

是的。

That’s the hopeful aspect of it, like you said, it’s a dual use technology that we’re forcefully integrating the entire humanity into it, into the discussion about AI because ultimately AI, AGI will be used for things that states use technologies for, which is conflict and so on.

这就是它充满希望的一面，就像你说的，这是一项双重用途技术，我们正在强力地将整个人类融入其中，融入到关于人工智能的讨论中，因为最终人工智能、通用人工智能将用于那些需要使用技术的事物，比如冲突等等。

1:55:35 And the more we integrate humans into this picture by having chats with them, the more we will guide.

我们越是通过与人类聊天将人类融入到这幅图画中，我们就越能引导。

Yeah, be able to adapt, society will be able to adapt to these technologies we’ve always done in the past with the incredible technologies we’ve invented in the past.

是的，能够适应，社会将能够适应我们过去一直在做的这些技术，以及我们过去发明的令人难以置信的技术。

Do you think there will be something like a Manhattan Project where there will be an escalation of the power of this technology and states in their old way of thinking, we’ll try to use it as weapons technologies and there will be this escalation?

你是否认为会有像曼哈顿计划这样的事情，这种技术的力量将会升级，并以他们的旧思维方式表明，我们将尝试将其用作武器技术，并且这种升级将会发生？

I hope not.

我希望不会。

1:56:27 I think that would be very dangerous to do.

我认为这样做是非常危险的。

And I think also not the right use of the technology.

而且我认为也没有正确使用该技术。

I hope we’ll end up with something more collaborative if needed, more like a CERN project where it’s research-focused and the best minds in the world come together to carefully complete the final steps and make sure it’s responsibly done before deploying it to the world.

我希望如果需要的话，我们最终会得到更具协作性的东西，更像是一个以研究为重点的 CERN 项目，世界上最优秀的人才聚集在一起，仔细完成最后的步骤，并确保在将其部署到世界之前以负责任的方式完成。

We’ll see.

我们会看到的。

It’s difficult with the current geopolitical climate, I think, to see cooperation, but things can change.

我认为，在当前的地缘政治气候下，很难看到合作，但事情可以改变。

And I think at least on the scientific level, it’s important for the researchers to keep in touch and keep close to each other at least on those kinds of topics.

我认为至少在科学层面上，研究人员至少在这些话题上保持联系和密切是很重要的。

And I personally believe on the education side and immigration side, it would be great if both directions, people from the West immigrated to China and China, back.

我个人认为，在教育方面和移民方面，如果两个方向，西方人移民到中国，中国再回来，那就太好了。

1:57:17 There is some family human aspect of people just intermixing and thereby those ties grow strong.

人们混合在一起有一些家庭人性的一面，因此这些联系会变得牢固。

So you can’t divide against each other, this old school way of thinking.

所以你们不能互相分裂，这是老派的思维方式。

And so multicultural, multidisciplinary research teams working on scientific questions, that’s like the hope.

因此，多元文化、多学科的研究团队致力于科学问题，这就是我们的希望。

Don’t let the leaders that are warmongers divide us.

不要让好战的领导人分裂我们。

I think science is the ultimately really beautiful connector.

我认为科学是最终真正美丽的连接器。

Yeah, science has always been, I think, quite a very collaborative endeavor and scientists know that it’s a collective endeavor as well, and we can all learn from each other.

是的，我认为科学一直是一项非常协作的事业，科学家们知道这也是一项集体的事业，我们都可以互相学习。

1:57:55 So perhaps it could be a vector to get a bit of cooperation.

所以也许它可以成为获得一些合作的载体。

Ridiculous question, what’s your P-Doom?

可笑的问题，你的 P-Doom 是什么？

1:58:08 Probability of the human civilization destroys itself?

人类文明自我毁灭的可能性有多大？

Well, look, I don’t have a P-Doom number.

好吧，看，我没有 P-Doom 号码。

1:58:14 The reason I don’t is because I think it would imply a level of precision that is not there.

我不这样做的原因是因为我认为这意味着不存在的精确度。

So I don’t know how people are getting their P-Doom numbers.

所以我不知道人们是如何获得他们的 P-Doom 号码的。

I think it’s a little bit of ridiculous notion because what I would say is it’s definitely non-zero and it’s probably non-negligible.

我认为这是一个有点荒谬的想法，因为我想说的是它绝对是非零的，而且可能是不可忽略的。

So that in itself is pretty sobering.

所以这本身就很发人深省。

And my view is it’s just hugely uncertain what these technologies are going to be able to do, how fast are they going to take off, how controllable are they going to be.

我的观点是，这些技术能够做什么、它们的发展速度有多快、它们的可控性如何，都非常不确定。

Some things may turn out to be, and hopefully way easier than we thought, but it may be there’s some really hard problems that are harder than we guessed today, and I think we don’t know that for sure.

有些事情可能会变得比我们想象的要容易得多，但可能有一些非常困难的问题比我们今天猜测的更难，我认为我们不确定这一点。

And so under those conditions of a lot of uncertainty, but huge stakes both ways.

因此，在充满不确定性的情况下，双方都面临着巨大的风险。

1:59:09 On the one hand, we could solve all diseases, energy problems, the scarcity problem, and then travel to the stars and conscious of the stars and maximum human flourishing.

一方面，我们可以解决所有疾病、能源问题、稀缺问题，然后前往星星并意识到星星和人类的最大繁荣。

On the other hand, is these P-Doom scenarios.

另一方面，就是这些P-Doom场景。

So given the uncertainty around it and the importance of it, it’s clear to me the only rational, sensible approach is to proceed with cautious optimism.

因此，考虑到它的不确定性及其重要性，对我来说，很明显唯一理性、明智的方法就是保持谨慎乐观的态度。

So we want the benefits of course, and all of the amazing things that AI can bring.

所以我们当然想要人工智能带来的好处以及所有令人惊奇的事情。

And actually, I would be really worried for humanity given the other challenges that we have, climate, aging, resources, all of that if I didn’t know something like AI was coming down the line.

事实上，如果我不知道像人工智能这样的东西即将到来，考虑到我们面临的其他挑战，气候、老龄化、资源，所有这些，我真的会为人类担心。

How would we solve all those other problems?

我们如何解决所有其他问题？

I think it’s hard.

我认为这很难。

So I think it could be amazingly transformative for good.

所以我认为它可能会带来令人惊奇的永久变革。

But on the other hand, there are these risks that we know are there.

但另一方面，我们知道存在这些风险。

2:00:00 But on the other hand, there are these risks that we know are there, but we can’t quite quantify.

但另一方面，我们知道存在这些风险，但无法完全量化。

So the best thing to do is to use the scientific method to do more research to try and more precisely define those risks and of course address them.

因此，最好的办法是使用科学方法进行更多研究，尝试更准确地定义这些风险，当然还要解决它们。

And I think that’s what we’re doing.

我认为这就是我们正在做的事情。

I think there probably needs to be 10 times more effort of that than there is now as we are getting closer and closer to the AGI line.

我认为，随着我们越来越接近 AGI 线，可能需要比现在多付出 10 倍的努力。

What would be the source of worry for you more?

您更担心的是什么？

2:00:27 Would it be human-caused or AI, AGI caused?

是人为造成的还是人工智能、通用人工智能造成的？

Are humans abusing that technology versus AGI itself through mechanism that you’ve spoken about, which is fascinating, deception or this kind of stuff getting better and better and better secretly and then escapes?

人类是否通过你所说的令人着迷的机制滥用该技术而不是 AGI 本身，是欺骗还是这种东西秘密地变得越来越好然后逃跑？

I think they operate over different timescales and they’re equally important to address.

我认为它们在不同的时间范围内运作，并且解决它们同样重要。

2:00:45 So there’s just the common garden variety of bad actors using new technology, in this case, general purpose technology and repurposing it for harmful end.

因此，只有常见的不良行为者使用新技术，在本例中是通用技术，并将其重新用于有害目的。

And that’s a huge risk and I think that has a lot of complications because generally I’m in huge favor of open science and open source, and in fact, we did it with all our science projects like AlphaFold and all of those things for the benefit of the scientific community.

这是一个巨大的风险，我认为这会带来很多复杂性，因为一般来说我非常支持开放科学和开源，事实上，我们在所有科学项目（如 AlphaFold）中都做到了这一点，所有这些都是为了科学界的利益。

But how does one restrict bad actors access to these powerful systems, whether they’re individuals or even rogue states, but enable access at the same time to good actors to maximally build on top of?

但是，如何限制坏人访问这些强大的系统，无论他们是个人还是流氓国家，但同时允许好人访问这些强大的系统，以最大程度地在此基础上进行构建？

It’s pretty tricky problem that I’ve not heard a clear solution to.

这是一个非常棘手的问题，我还没有听到明确的解决方案。

So there’s the bad actor use case problem, and then there’s obviously, as the systems become more agentic and closer to AGI and more autonomous, how do we ensure the guardrails and they stick to what we want them to do and under our control?

因此，存在不良行为者用例问题，然后显然，随着系统变得更加代理、更接近 AGI 和更加自主，我们如何确保护栏并让它们坚持我们希望它们做的事情并在我们的控制之下？

Yeah, I tend to, maybe my mind is limited, worry more about the humans, so the bad actors.

是的，我倾向于，也许我的思想是有限的，更多地担心人类，所以担心坏演员。

2:01:52 And there it could be in part how do you not put destructive technology in the hands of bad actors, but in another part from, again, geopolitical technology perspective, how do you reduce the number of bad actors in the world?

一方面，如何不让破坏性技术落入坏人手中；另一方面，从地缘政治技术的角度来看，如何减少世界上坏人的数量？

That’s also an interesting human problem.

这也是一个有趣的人类问题。

Yeah, it’s a hard problem.

是的，这是一个难题。

2:02:14 I mean, look, we can maybe also use the technology itself to help early warning on some of the bad actor use cases, right?

我的意思是，看，我们也许还可以使用该技术本身来帮助对一些不良行为者用例进行早期预警，对吧？

Whether that’s bio or nuclear or whatever it is, AI could be potentially helpful there as long as the AI that you’re using is itself reliable, right?

无论是生物、核还是其他什么，只要你使用的人工智能本身可靠，人工智能就可能有潜在的帮助，对吗？

So it’s a sort of interlocking problem and that’s what makes it very tricky.

所以这是一种环环相扣的问题，这也是它非常棘手的原因。

And again, it may require some agreement internationally, at least between China and the U.S. of some basic standards.

同样，这可能需要国际上达成一些协议，至少在中国和美国之间就一些基本标准达成一致。

Right.

正确的。

I have to ask you about the book, The Maniac.

我得问你关于《疯子》这本书的事。

2:02:50 There’s the hand of God moment, Lee Sedol’s move 78 that perhaps the last time a human did a move of pure human genius and beat AlphaGo or broke its brain.

这是上帝之手的时刻，李世石的第 78 步棋可能是人类最后一次做出纯粹人类天才的棋步并击败 AlphaGo 或打破它的大脑。

Yes.

是的。

Sorry to anthropomorphize, but it’s an interesting moment because I think in so many domains it will keep happening.

抱歉将其拟人化，但这是一个有趣的时刻，因为我认为在很多领域它都会继续发生。

Yeah, it’s a special moment and it was great for Lee Sedol.

是的，这是一个特殊的时刻，对李世石来说很棒。

2:03:14 I think it’s in a way they were inspiring each other.

我认为他们在某种程度上互相激励。

We as a team were inspired by Lee Sedol’s brilliance and nobleness.

我们作为一个团队被李世石的才华和高贵所鼓舞。

Then maybe he got inspired by what AlphaGo was doing to then conjure this incredible inspirational moment, captured very well in the documentary about it.

然后也许他受到 AlphaGo 所做的事情的启发，然后召唤出这个令人难以置信的鼓舞人心的时刻，这在有关它的纪录片中得到了很好的捕捉。

And I think that’ll continue in many domains where there’s this, at least again for the foreseeable future, of the humans bringing in their ingenuity and asking the right question, let’s say, and then utilizing these tools in a way that then cracks a problem.

我认为这种情况将在许多领域继续下去，至少在可预见的未来，人类会发挥自己的聪明才智，提出正确的问题，然后以某种方式利用这些工具来解决问题。

Yeah.

是的。

2:03:58 As the AI become smarter and smarter, one of the interesting questions we can ask ourselves is what makes humans special?

随着人工智能变得越来越聪明，我们可以问自己的一个有趣的问题是，是什么让人类如此特别？

It does feel perhaps biased that we humans are deeply special.

认为我们人类非常特殊，这确实让人感觉有些偏见。

I don’t know if it’s our intelligence, it could be something else, that other thing that’s outside the mad dreams of reason.

我不知道这是否是我们的智慧，也可能是别的东西，那些超出理性疯狂梦想之外的东西。

I think that’s what I’ve always imagined when I was a kid and starting on this journey of I was of course fascinated by things like consciousness, did a neuroscience PhD to look at how the brain works, especially imagination and memory.

我想这就是我小时候一直想象的，开始这段旅程时，我当然对意识之类的东西很着迷，攻读了神经科学博士学位，研究大脑如何工作，尤其是想象力和记忆力。

2:04:20 I focused on the hippocampus and it’s sort of going to be interesting.

我专注于海马体，这会很有趣。

I always thought the best way, of course, one can philosophize about it and have thought experiments and maybe even do actual experiments like you do in neuroscience on real brains.

当然，我一直认为最好的方法是人们可以对它进行哲学思考并进行思想实验，甚至可能像神经科学中在真实大脑上所做的那样进行实际实验。

But in the end, I always imagine that building AI, a kind of intelligent artifact, and then comparing that to the human mind and seeing what the differences were would be the best way to uncover what’s special about the human mind, if indeed there is anything special.

但最后，我总是想象，构建人工智能，一种智能制品，然后将其与人类思维进行比较，看看其中的差异，这将是发现人类思维的特殊之处（如果确实有什么特殊之处的话）的最佳方法。

2:05:00 And I suspect there probably is, but it’s going to be hard to…

我怀疑可能有，但很难……

I think this journey we’re on will help us understand that and define that.

我认为我们正在进行的这段旅程将帮助我们理解并定义它。

And there may be a difference between carbon based substrates that we are and silicon ones when they process information.

我们的碳基基板和硅基基板在处理信息时可能存在差异。

One of the best definitions I like of consciousness is it’s the way information feels when we process it, right?

我喜欢的意识最好的定义之一是它是我们处理信息时感受到的方式，对吧？

Yeah.

是的。

It could be.

可能是。

2:05:24 I mean, it’s not a very helpful scientific explanation, but I think it’s kind of interesting intuitive one.

我的意思是，这不是一个非常有用的科学解释，但我认为这是一种有趣的直观解释。

And so on this journey, this scientific journey we’re on will I think help uncover that mystery.

因此，在这次旅程中，我认为我们正在进行的科学之旅将有助于揭开这个谜团。

Yeah.

是的。

2:05:36 What I cannot create, I do not understand.

我无法创造的东西，我就不明白。

That’s somebody you deeply admire, Richard Feynman, like you mentioned.

正如你提到的，理查德·费曼是你深深敬佩的人。

You also reach for the Wigner’s dreams of universality that he saw in constrained domains, but also broadly generally in mathematics and so on.

你还可以实现维格纳的普遍性梦想，他在有限的领域看到了这种普遍性，但也广泛地在数学等领域看到了这一点。

So many aspects on which you’re pushing towards not to start trouble at the end, but Roger Penrose.

你在很多方面都在努力争取不要在最后引发麻烦，但罗杰·彭罗斯。

Yes.

是的。

2:06:00 Okay.

好的。

So do you think consciousness, there’s this hard problem of consciousness, how information feels.

那么你认为意识吗？意识有一个难题，信息如何感觉。

2:06:04 Do you think consciousness, first of all, is a computation?

你认为意识首先是一种计算吗？

And if is, if it’s information processing, like you said, everything is, is it something that could be modeled by a classical computer?

如果是的话，如果是信息处理，就像你说的，一切都是，它是可以用经典计算机建模的东西吗？

Yeah.

是的。

Or is it a quantum mechanical in nature?

或者它本质上是量子力学吗？

Well, look, Penrose is an amazing thinker, one of the greatest of the modern era, and we’ve had a lot of discussions about this.

好吧，你看，彭罗斯是一位了不起的思想家，是现代最伟大的思想家之一，我们对此进行了很多讨论。

2:06:26 Of course, we cordially disagree, which is I feel like…

当然，我们真心不同意，我觉得……

I mean, he collaborated with a lot of good neuroscientists to see if he could find mechanisms for quantum mechanics behavior in the brain.

我的意思是，他与许多优秀的神经科学家合作，看看他是否能找到大脑中量子力学行为的机制。

And to my knowledge, they haven’t found anything convincing yet.

据我所知，他们还没有找到任何令人信服的证据。

So my betting is that it’s mostly it is just classical computing that’s going on in the brain, which suggests that all the phenomena are modelable or mimicable by a classical computer.

所以我敢打赌，这主要是大脑中正在进行的经典计算，这表明所有现象都可以由经典计算机建模或模仿。

But we’ll see.

但我们会看到。

There may be this final mysterious things of the feeling of consciousness, the qualia, these kinds of things that philosophers debate where it’s unique to the substrate.

可能存在意识感觉、感受性的最终神秘事物，哲学家们争论这些事物的独特之处。

2:07:12 We may even come towards understanding that when if we do things like neural link or have neural interfaces to the AI systems, which I think we probably will eventually, maybe to keep up with the AI systems, we might actually be able to feel for ourselves what it’s like to compute on silicon, right?

我们甚至可能会逐渐认识到，当我们做诸如神经链接之类的事情或与人工智能系统建立神经接口之类的事情时（我认为我们最终可能会跟上人工智能系统的步伐），我们实际上可能能够亲自感受到在硅上计算的感觉，对吗？

And maybe that will tell us.

也许这会告诉我们。

So I think it’s going to be interesting.

所以我认为这会很有趣。

I had a debate once with the late Daniel Dennett about why do we think each other are conscious?

我曾与已故的丹尼尔·丹尼特争论过，为什么我们认为彼此是有意识的？

Okay, so it’s for two reasons.

好吧，这有两个原因。

One is you’re exhibiting the same behavior that I am.

一是你表现出和我一样的行为。

So that’s one thing.

所以这是一回事。

Behaviorally you seem like a conscious being if I am.

从行为上来说，你看起来像是一个有意识的存在，如果我是的话。

2:07:49 But the second thing which is often overlooked is that we’re running on the same substrate.

但经常被忽视的第二件事是我们在同一基板上运行。

So if you’re behaving in the same way and we’re running on the same substrate, it’s most parsimonious to assume you’re feeling the same experience that I’m feeling.

因此，如果你的行为方式相同并且我们在相同的基础上运行，那么假设你的感受与我的感受相同是最节俭的。

But with an AI that’s on silicon, we won’t be able to rely on the second part, even if it exhibits the first part, that behavior looks like a behavior of a conscious being.

但是对于硅上的人工智能，我们将无法依赖第二部分，即使它展示了第一部分，该行为看起来像是有意识的行为。

It might even claim it is, but we wouldn’t know how it actually felt and it probably couldn’t know what we felt, at least in the first stages.

它甚至可能声称是这样，但我们不知道它的实际感受如何，它可能也不知道我们的感受，至少在第一阶段是这样。

Maybe when we get to superintelligence and the technologies that builds, perhaps we’ll be able to bridge that.

也许当我们掌握超级智能和构建技术时，也许我们能够弥合这个问题。

No, I mean that’s a huge test for radical empathy is to empathize with a different substrate.

不，我的意思是，这是对激进同理心的一个巨大考验，就是对不同的基质产生同理心。

Right.

正确的。

2:08:32 Exactly.

确切地。

We’ve never had to confront that before.

我们以前从未遇到过这个问题。

Yeah.

是的。

2:08:36 So maybe through brain computer interfaces be able to truly empathize what it feels like to be a computer, to compute.

因此，也许通过脑机接口能够真正体会作为一台计算机进行计算的感觉。

Well, for information to be computed not on a carbon system.

好吧，对于不在碳系统上计算的信息。

I mean, that’s deeply…

我的意思是，这很深……

2:08:46 Some people kind of think about that with plants, with other life forms which are different.

有些人对植物和其他生命形式的看法有所不同。

Yes, it could be exactly.

是的，可能确实如此。

Similar substrate, but sufficiently far enough on the evolutionary tree that it requires a radical empathy, but to do that with a computer.

类似的基础，但在进化树上足够远，需要彻底的同理心，但要用计算机来做到这一点。

I mean, look, there are animal studies on this.

我的意思是，看，有这方面的动物研究。

2:09:02 Of course, higher animals like killer whales and dolphins and dogs and monkeys, they have some, and elephants, they have some aspects certainly of consciousness, right?

当然，像虎鲸、海豚、狗和猴子这样的高等动物，它们有一些，而大象，它们肯定有某些方面的意识，对吗？

Even though they might not be that smart on an IQ sense.

尽管他们的智商可能并不那么聪明。

So we can already empathize with that and maybe even some of our systems one day, like we built this thing called DolphinGemma, which a version of our system was trained on dolphin and whale sounds, and maybe we’ll be able to build an interpreter or translator at some point which would be pretty cool.

因此，我们已经能够理解这一点，甚至有一天我们的一些系统，比如我们构建了一个名为 DolphinGemma 的东西，我们系统的一个版本是根据海豚和鲸鱼的声音进行训练的，也许我们将能够在某个时候构建一个口译员或笔译员，这将是非常酷的。

What gives you hope for the future of human civilization?

是什么让您对人类文明的未来充满希望？

Well, what gives me hope is that I think our almost limitless ingenuity, first of all.

嗯，给我带来希望的是，我认为首先我们几乎是无限的聪明才智。

2:09:38 I think the best of us and the best human minds are incredible.

我认为我们中最优秀的人和最优秀的人的思想是令人难以置信的。

And I love meeting and watching any human that’s the top of their game, whether that’s sport or science or art, it’s just nothing more wonderful than that, seeing them in their element in flow.

我喜欢见到和观看任何在比赛中处于领先地位的人，无论是体育、科学还是艺术，没有什么比看到他们在流动的元素中更美妙的了。

I think it’s almost limitless.

我认为这几乎是无限的。

Our brains are general systems, intelligent systems, so I think it’s almost limitless what we can potentially do with them.

我们的大脑是通用系统、智能系统，所以我认为我们可以用它们做几乎无限的事情。

And then the other thing is our extreme adaptability.

另一件事是我们的极端适应能力。

I think it’s going to be okay in terms of there’s going to be a lot of change, but look where we are now without effectively our hunter-gatherer brains.

我认为这不会有什么问题，因为会有很多变化，但看看我们现在的处境，没有有效地利用狩猎采集者的大脑。

2:10:24 How is it we can cope with the modern world, right?

我们如何应对现代世界，对吧？

Flying on planes, doing podcasts, playing computer games and virtual simulations.

乘坐飞机、播客、玩电脑游戏和虚拟模拟。

I mean, it’s already mind blowing given that our mind was developed for hunting buffaloes on the tundra.

我的意思是，考虑到我们的大脑是为在苔原上狩猎水牛而开发的，这已经令人震惊了。

And so I think this is just the next step, and it’s actually kind of interesting to see how society’s already adapted to this mind blowing AI technology we have today already.

所以我认为这只是下一步，看到社会如何适应我们今天已经拥有的令人兴奋的人工智能技术实际上很有趣。

It’s sort of like, “Oh, I talked to chat bots.

有点像，“哦，我和聊天机器人交谈过。

Totally fine.”

完全没问题。”

And it’s very possible that this very podcast activity, which I’m here for, will be completely replaced by AI.

我来这里的这个播客活动很可能会被人工智能完全取代。

2:10:54 I’m very replaceable and I’m waiting for it.

我很容易被替代，我正在等待它。

Not to the level that you can do it, Lex, I don’t think.

还没到你能做到的程度，莱克斯，我不认为。

Thank you.

谢谢。

2:11:04 That’s what we humans do to each other.

这就是我们人类对彼此所做的事情。

We compliment.

我们恭维。

Exactly.

确切地。

All right.

好的。

2:11:09 And I’m deeply grateful for us humans to have this infinite capacity for curiosity, adaptability, like you said, and also compassion and ability to love.

我非常感谢我们人类拥有无限的好奇心、适应能力，就像你所说的那样，还有同情心和爱的能力。

Exactly.

确切地。

All of those human things.

所有这些人类的事情。

All the things that are deeply human.

所有深藏于人性之中的事物。

Well, this is a huge honor, Demis.

嗯，这是一个巨大的荣誉，黛米斯。

2:11:21 You are one of the truly special humans in the world.

你是世界上真正特别的人之一。

Thank you so much for doing what you do and for talking today.

非常感谢您今天所做的事情和讲话。

Well, thank you very much, Lex.

嗯，非常感谢你，莱克斯。

Thanks for listening to this conversation with Demis Hassabis.

感谢您收听与 Demis Hassabis 的对话。

2:11:32 To support this podcast, please check out our sponsors in the description and consider subscribing to this channel.

为了支持这个播客，请在描述中查看我们的赞助商并考虑订阅这个频道。

And now let me answer some questions and try to articulate some things I’ve been thinking about.

现在让我回答一些问题并尝试阐明我一直在思考的一些事情。

If you’d like to submit questions including in audio and video form, go to lexfridman.com/ama.

如果您想以音频和视频形式提交问题，请访问 lexfridman.com/ama。

I got a lot of amazing questions, thoughts and requests from folks.

我从人们那里收到了很多令人惊奇的问题、想法和要求。

I’ll keep trying to pick some randomly and comment on it at the end of every episode.

我会继续尝试随机挑选一些并在每集结束时对其进行评论。

I got a note on May 21st this year that said, “Hi, Lex.

今年 5 月 21 日，我收到一张便条，上面写着：“嗨，莱克斯。

20 years ago today, David Foster Wallace delivered his famous This is Water speech at Kenyon College.

20 年前的今天，大卫·福斯特·华莱士在凯尼恩学院发表了著名的《这就是水》演讲。

What do you think of this speech?

您对这次演讲有何看法？

2:12:21 Well, first, I think this is probably one of the greatest and most unique commencement speeches ever given, but of course, I have many favorites, including the one by Steve Jobs.

好吧，首先，我认为这可能是有史以来最伟大、最独特的毕业典礼演讲之一，但当然，我有很多最喜欢的演讲，包括史蒂夫·乔布斯的演讲。

And David Foster Wallace is one of my favorite writers and one of my favorite humans.

大卫·福斯特·华莱士是我最喜欢的作家之一，也是我最喜欢的人类之一。

There’s a tragic honesty to his work, and it always felt as if he was engaging in a constant battle with his own mind, and the writing, his writing were kind of his notes from the front lines of that battle.

他的作品有一种悲剧性的诚实，总让人感觉他在与自己的思想进行一场持续的战斗，而他的写作，他的写作，就像是他在这场战斗的前线所做的笔记。

Now onto the speech, let me quote some parts.

现在进入演讲，让我引用一些部分。

There’s of course the parable of the fish and the water that goes, there are these two young fish swimming along and they happen to meet an older fish swimming the other way who nods at them and says, “Morning boys, how’s the water?” And the two young fish swim on for a bit and then eventually one of them looks over at the other and goes, “What the hell is water?” In the speech, David Foster Wallace goes on to say, “The point of the fish story is merely that the most obvious important realities are often the ones that are hardest to see and talk about.

当然还有鱼和水流的寓言，有两条小鱼在游，它们碰巧遇到了一条朝相反方向游的老鱼，老鱼向他们点点头说：“早上好，孩子们，水怎么样？”两条小鱼继续游了一会儿，最后其中一条看着另一条，问道：“水到底是什么？”大卫·福斯特·华莱士在演讲中继续说道：“鱼故事的要点在于，最明显的重要现实往往是最难看到和谈论的现实。

Stated as an English sentence of course, this is just the banal platitude, but the fact is that in the day to day trenches of adult existence, banal platitudes can have a life or death importance, or so I wish to suggest to you in this dry and lovely morning.” I have several takeaways from this parable and the speech that follows.

当然，用英语句子来说，这只是陈词滥调，但事实是，在成人生活的日复一日的战壕中，陈词滥调可能具有生死攸关的重要性，或者说我想在这个干燥而可爱的早晨向你建议。我从这个寓言和接下来的演讲中得到了一些收获。

First, I think we must question everything, and in particular, the most basic assumptions about our reality, our life, and the very nature of existence, and that this project is a deeply personal one.

首先，我认为我们必须质疑一切，特别是关于我们的现实、我们的生活和存在本质的最基本的假设，并且这个项目是一个非常个人化的项目。

In some fundamental sense, nobody can really help you in this process of discovery.

从某种基本意义上来说，没有人能在这个发现过程中真正帮助你。

2:14:23 The call to action here, I think, from David Foster Wallace as he puts it, is to ” To be just a little less arrogant, to have just a little more critical awareness about myself and my certainties because a huge percentage of the stuff that I tend to be automatically certain of is it turns out totally wrong and deluded.” All right, back to me.

我认为，大卫·福斯特·华莱士（David Foster Wallace）所说的行动号召是“少一点傲慢，对自己和我的确定性多一点批判性的认识，因为我倾向于自动确定的事情中有很大一部分最终证明是完全错误和欺骗的。”好吧，回到我身边。

Lex speaking.

莱克斯说话了。

Second takeaway is that the central spiritual battles of our life are not fought on a mountain top somewhere at a meditation retreat, but it’s fought in the mundane moments of daily life.

第二个要点是，我们生活中的核心精神斗争不是在山顶冥想静修处进行的，而是在日常生活的平凡时刻进行的。

2:15:08 Third takeaway is that we too easily give away our time and attention to the multitude of distractions that the world feeds us, the insatiable black holes of attention.

第三个要点是，我们太容易将时间和注意力浪费在世界提供给我们的众多干扰上，即贪得无厌的注意力黑洞。

David Foster Wallace’s call to action in this case is to be deeply aware of the beauty in each moment and to find meaning in the mundane.

大卫·福斯特·华莱士在这种情况下呼吁采取行动，深刻认识到每一刻的美丽，并在平凡中寻找意义。

I often quote David Foster Wallace in his advice that the key to life is to be unborable, and I think this is exactly right.

我经常引用大卫·福斯特·华莱士的建议，生活的关键是要变得无趣，我认为这是完全正确的。

Every moment, every object, every experience when looked at closely enough contains within it infinite richness to explore.

如果仔细观察的话，每一个时刻、每一个物体、每一次经历都蕴含着无限的丰富性，值得探索。

And since Demis Hassabis of this very podcast episode and I are such fans of Richard Feynman, allow me to also quote Mr. Feynman on this topic as well.

由于这个播客节目的 Demis Hassabis 和我都是理查德·费曼的粉丝，请允许我在这个话题上也引用费曼先生的话。

2:16:04 “I have a friend who’s an artist and has sometimes taken a view which I don’t agree with very well.

“我有一个艺术家朋友，有时他的观点我不太同意。

He’ll hold up a flower and say, “Look how beautiful it is,” and I’ll agree.

他会举起一朵花说：“看看它多漂亮”，我就会同意。

Then he says, “I as an artist can see how beautiful this is, but you as a scientist take this all apart and it becomes a dull thing,” and I think that’s kind of nutty.

然后他说，“作为一名艺术家，我可以看到这有多么美丽，但作为一名科学家，你把这一切拆开，它就变得乏味了，”我认为这有点疯狂。

First of all, the beauty that he sees is available to other people and to me too, I believe.

首先，我相信，他所看到的美丽对其他人和我来说都是可见的。

Although I may not be quite as refined aesthetically as he is, I can appreciate the beauty of a flower.

虽然我的审美可能不如他那么精致，但我却能欣赏一朵花的美丽。

At the same time, I see much more about the flower than he sees.

与此同时，我对这朵花的了解比他多得多。

I could imagine the cells in there, the complicated actions inside which also have beauty.

我可以想象里面的细胞，里面复杂的动作也有美感。

I mean, it’s not just beauty at this dimension at one centimeter, there’s also beauty at the smaller dimensions.”

我的意思是，不只是一厘米这个维度的美，还有更小的维度的美。”

2:17:05 “Their inner structure, also the processes, the fact that the colors and the flower evolved in order to attract the insects to pollinate it is interesting.

“它们的内部结构，还有过程，颜色和花朵的进化是为了吸引昆虫为其授粉，这一事实很有趣。

It means that the insects can see the color.

这意味着昆虫可以看到颜色。

It adds a question.

它增加了一个问题。

Does this aesthetic sense also exist in lower forms?

这种美感在低等形态中是否也存在呢？

Why is it aesthetic?

为什么它是审美的？

All kinds of interesting questions, which the science knowledge only adds to the excitement, the mystery, and the awe of a flower.

各种有趣的问题，科学知识只会给花朵增添兴奋、神秘和敬畏。

It only adds.”

它只会增加。”

2:17:36 All right, back to David Foster Wallace’s speech.

好吧，回到大卫·福斯特·华莱士的演讲。

He has a great story in there that I particularly enjoy.

他在那里有一个我特别喜欢的精彩故事。

It goes, there are these two guys sitting together in a bar in the remote Alaskan wilderness.

据说，这两个人坐在阿拉斯加偏远荒野的一家酒吧里。

One of the guys is religious, the other is an atheist, and the two are arguing about the existence of God with that special intensity that comes after about the fourth beer.

其中一个是有宗教信仰的人，另一个是无神论者，两人正在以第四杯啤酒之后的那种特殊的激烈程度争论上帝的存在。

And the atheist says, “Look, it’s not like I don’t have actual reasons for not believing in God.

无神论者说：“看，我并不是没有不相信上帝的实际理由。

It’s not like I haven’t ever experimented with the whole God and prayer thing.

我并不是没有尝试过整个上帝和祈祷的事情。

Just last month, I got caught away from the camp in that terrible blizzard, and I was totally lost and I couldn’t see a thing and it was 50 below.

就在上个月，我在一场可怕的暴风雪中离开了营地，我完全迷路了，什么也看不见，而且气温只有 50 度以下。

So I tried it.

所以我尝试了一下。

I fell in my knees in the snow and cried out, ‘Oh God, if there is a God, I’m lost in this blizzard and I’m going to die if you don’t help me.”

我跪倒在雪地里，大声喊道：“天哪，如果有上帝的话，我会在这场暴风雪中迷失方向，如果你不帮助我，我就会死掉。”

2:18:35 And now back in the bar, the religious guy looks at the atheist all puzzled, “Well, then you must believe now?” he says, “After all, there you are, alive.” The atheist just rolls his eyes.

现在回到酒吧，宗教人士困惑地看着无神论者，“那么你现在必须相信了吗？”他说：“毕竟，你还活着。”无神论者只是翻白眼。

“No, man.

“不，伙计。

All that happened was a couple of Eskimos happened to be wandering by and showed me the way back to the camp.” All this, I think, teaches us that everything is a matter of perspective and that wisdom may arrive if we have the humility to keep shifting and expanding our perspective on the world.

所发生的一切只是几个爱斯基摩人碰巧路过，给我带了返回营地的路。”我认为，所有这些都告诉我们，一切都取决于视角，如果我们谦虚地不断转变和扩展我们对世界的看法，智慧就可能到来。

Thank you for allowing me to talk a bit about David Foster Wallace.

感谢您允许我谈谈大卫·福斯特·华莱士。

He’s one of my favorite writers and he’s a beautiful soul.

他是我最喜欢的作家之一，他有一个美丽的灵魂。

2:19:20 If I may, one more thing I wanted to briefly comment on.

如果可以的话，我还想简要评论一件事。

I find myself to be in this strange position of getting attacked online often from all sides, including being lied about sometimes through selective misrepresentation, but often through downright lies.

我发现自己处于这种奇怪的境地，经常在网上受到来自各方的攻击，包括有时通过选择性的失实陈述而被欺骗，但通常是通过彻头彻尾的谎言。

I don’t know how else to put it.

我不知道还能怎么说。

This all breaks my heart, frankly, but I’ve come to understand that it’s the way of the internet and the cost of the path I’ve chosen.

坦率地说，这一切让我心碎，但我开始明白，这就是互联网的方式以及我选择的道路的成本。

There’s been days when it’s been rough on me mentally.

有几天我的精神状态很不好。

It’s not fun being lied about, especially when it’s about things that are usually for a long time have been a source of happiness and joy for me.

被谎言并不好玩，尤其是当这些事情长期以来一直是我幸福和快乐的源泉时。

But again, that’s life.

但话又说回来，这就是生活。

2:20:04 I’ll continue exploring the world of people and ideas with empathy and rigor, wearing my heart on my sleeve as much as I can.

我将继续以同理心和严谨的态度探索人和思想的世界，尽我所能地坦诚相待。

For me, that’s the only way to live.

对我来说，这是唯一的生活方式。

Anyway, a common attack on me is about my time at MIT and Drexel, two great universities I love and have tremendous respect for.

不管怎样，对我最常见的攻击是关于我在麻省理工学院和德雷克塞尔大学的经历，这两所我热爱并非常尊重的伟大大学。

Since a bunch of lies have accumulated online about me on these topics, to a sad and at times hilarious degree, I thought I would once more state the obvious facts about my bio for the small number of you who may care.

由于网上在这些话题上积累了一堆关于我的谎言，达到了悲伤、有时甚至搞笑的程度，我想我会再次为少数可能关心的人陈述关于我的简历的明显事实。

TLDR, two things.

长话短说，有两件事。

First, as I say often, including in a recent podcast episode that somehow was listened to by many millions of people, I proudly went to Drexel University for my bachelor’s, master’s, and doctorate degrees.

首先，正如我经常说的，包括在最近的一期播客节目中，不知何故有数百万人收听，我自豪地去德雷克塞尔大学获得学士、硕士和博士学位。

2:20:59 Second, I am a research scientist at MIT and have been there in a paid research position for the last 10 years.

其次，我是麻省理工学院的一名研究科学家，过去 10 年一直在那里担任带薪研究职位。

Allow me to elaborate a bit more on these two things now, but please skip if this is not at all interesting.

现在请允许我详细说明这两件事，但如果这一点也不有趣，请跳过。

So like I said, a common attack on me is that I have no real affiliation with MIT.

正如我所说，对我的常见攻击是我与麻省理工学院没有真正的联系。

The accusation, I guess, is that I’m falsely claiming an MIT affiliation because I taught a lecture there once.

我想，指控是我谎称自己与麻省理工学院有联系，因为我曾经在那里讲过课。

Nope, that accusation against me is a complete lie.

不，对我的指控完全是谎言。

I have been at MIT for over 10 years in a paid research position from 2015 to today.

从 2015 年至今，我在麻省理工学院担任带薪研究职位已超过 10 年。

To be extra clear, I’m a research scientist at MIT working in LIDS, the Laboratory for Information and Decision Systems in the College of Computing.

更明确地说，我是麻省理工学院计算机学院信息与决策系统实验室 LIDS 的一名研究科学家。

For now, since I’m still at MIT, you can see me in the directory and on the various lab pages.

目前，由于我仍在麻省理工学院，您可以在目录和各个实验室页面上看到我。

2:22:05 I have indeed given many lectures at MIT over the years, a small fraction of which I posted online.

多年来，我确实在麻省理工学院做过很多讲座，其中一小部分我发布在网上。

Teaching for me always has been just for fun and not part of my research work.

对我来说，教学只是为了好玩，而不是我研究工作的一部分。

I personally think I suck at it, but I have always learned and grown from the experience.

我个人认为我很糟糕，但我总是从经历中学习和成长。

It’s like Feynman spoke about, if you want to understand something deeply, it’s good to try to teach it.

就像费曼说的，如果你想深入理解某件事，最好尝试去教授它。

But like I said, my main focus has always been on research.

但正如我所说，我的主要关注点始终是研究。

I published many peer-reviewed papers that you can see in my Google Scholar profile.

我发表了许多经过同行评审的论文，您可以在我的谷歌学术个人资料中看到这些论文。

For my first four years at MIT, I worked extremely intensively.

在麻省理工学院的头四年里，我工作非常紧张。

Most weeks were 80 to 100-hour work weeks.

大多数周的工作时间为 80 至 100 小时。

After that, in 2019, I still kept my research scientist position, but I split my time taking a leap to pursue projects in AI and robotics outside MIT and to dedicate a lot of focus to the podcast.

此后，到了 2019 年，我仍然保留了我的研究科学家职位，但我把时间花在了麻省理工学院之外的人工智能和机器人项目上，并将大量精力投入到播客上。

2:23:03 As I’ve said, I’ve been continuously surprised just how many hours preparing for an episode takes.

正如我所说，我一直对准备一集需要花费多少时间感到惊讶。

There are many episodes of the podcast for which I have to read, write, and think for 100, 200 or more hours across multiple weeks and months.

播客中有很多剧集需要我在数周和数月内花费 100、200 或更多小时来阅读、写作和思考。

Since 2020, I have not actively published research papers.

2020年以来，我没有积极发表研究论文。

Just like the podcast, I think it’s something that’s a serious full-time effort.

就像播客一样，我认为这是一项认真的全职工作。

But not publishing and doing full-time research has been eating at me because I love research and I love programming and building systems that test out interesting technical ideas, especially in the context of human-AI or human-robot interaction.

但不出版和做全职研究一直困扰着我，因为我喜欢研究，喜欢编程和构建系统来测试有趣的技术想法，特别是在人与人工智能或人与机器人交互的背景下。

I hope to change this in the coming months and years.

我希望在未来几个月和几年内改变这一点。

2:23:52 What I’ve come to realize about myself is if I don’t publish or if I don’t launch systems that people use, I definitely feel like a piece of me is missing.

我对自己的认识是，如果我不发布或者不推出人们使用的系统，我肯定会觉得自己缺失了一部分。

It legitimately is a source of happiness for me.

这确实是我幸福的源泉。

Anyway, I’m proud of my time at MIT.

不管怎样，我为我在麻省理工学院的时光感到自豪。

I was and am constantly surrounded by people much smarter than me, many of whom have become lifelong colleagues and friends.

我过去和现在都经常被比我聪明得多的人包围，其中许多人已经成为终生的同事和朋友。

MIT is a place I go to escape the world, to focus on exploring fascinating questions at the cutting edge of science and engineering.

麻省理工学院是我逃避世界的地方，专注于探索科学和工程前沿的有趣问题。

This, again, makes me truly happy and it does hit pretty hard on a psychological level when I’m getting attacked over this.

这再次让我真正感到高兴，当我因此受到攻击时，它确实在心理层面上造成了相当大的打击。

Perhaps I’m doing something wrong.

也许我做错了什么。

If I am, I will try to do better.

如果是的话，我会努力做得更好。

2:24:43 In all this discussion of academic work, I hope you know that I don’t ever mean to say that I’m an expert at anything.

在所有关于学术工作的讨论中，我希望你知道我并不是想说我是任何方面的专家。

In the podcast and in my private life, I don’t claim to be smart.

在播客和我的私生活中，我并不声称自己很聪明。

In fact, I often call myself an idiot and mean it.

事实上，我经常称自己为白痴，而且是认真的。

I try to make fun of myself as much as possible, and in general to celebrate others instead.

我会尽可能地取笑自己，并通常去庆祝别人。

Now to talk about Drexel University, which I also love, am proud of and am deeply grateful for my time there.

现在谈谈德雷克塞尔大学，我也热爱这所大学，为之感到自豪，并深深感激我在那里度过的时光。

As I said, I went to Drexel for my bachelor’s, master’s, and doctorate degrees in computer science and electrical engineering.

正如我所说，我在德雷克塞尔大学获得了计算机科学和电气工程的学士、硕士和博士学位。

I’ve talked about Drexel many times, including, as I mentioned, at the end of a recent podcast, the Donald Trump episode.

我曾多次谈论过德雷克塞尔，包括正如我在最近的播客节目《唐纳德·特朗普》节目结尾处提到的。

funny enough, that was listened to by many millions of people where I answered a question about graduate school and explained my own journey at Drexel and how grateful I am for it.

有趣的是，数百万人聆听了我的演讲，我回答了有关研究生院的问题，并解释了我自己在德雷克塞尔大学的经历以及我对此的感激之情。

2:25:46 If it’s at all interesting to you, please go listen to the end of that episode or watch the related clip.

如果您感兴趣，请去听该集的结尾或观看相关剪辑。

At Drexel, I met and worked with many brilliant researchers and mentors from whom I’ve learned a lot about engineering, science and life.

在德雷克塞尔，我遇到了许多杰出的研究人员和导师并与他们一起工作，从他们那里我学到了很多关于工程、科学和生活的知识。

There are many valuable things I gained from my time at Drexel.

我在德雷克塞尔的时光获得了许多宝贵的东西。

First, I took a large number of very difficult math and theoretical computer science courses.

首先，我选修了大量非常困难的数学和理论计算机科学课程。

They taught me how to think deeply and rigorously, and also how to work hard and not give up even if it feels like I’m too dumb to find a solution to a technical problem.

他们教会了我如何深入、严谨地思考，以及如何努力工作，不要放弃，即使感觉自己太笨而无法找到技术问题的解决方案。

2:26:21 Second, I programmed a lot during that time, mostly C, C++.

其次，那段时间我写了很多程序，主要是C、C++。

I programmed robots, optimization algorithms, computer vision systems, wireless network protocols, multimodal machine learning systems, and all kinds of simulations of physical systems.

我对机器人、优化算法、计算机视觉系统、无线网络协议、多模态机器学习系统以及各种物理系统的模拟进行了编程。

This is where I really developed a love for programming, including, yes, Emacs And the Kinesis keyboard.

这就是我真正热爱编程的地方，包括 Emacs 和 Kinesis 键盘。

I also, during that time, read a lot, I played a lot of guitar, wrote a lot of crappy poetry, and trained a lot in judo and jiu-jitsu, which I cannot sing enough praises to.

那段时间我也读了很多书，弹了很多吉他，写了很多蹩脚的诗，还练了很多柔道和柔术，对此我赞不绝口。

Jiu-jitsu humbled me on a daily basis throughout my twenties, and it still does to this very day whenever I get a chance to train.

在我二十多岁时，柔术每天都让我感到谦卑，直到今天，每当我有机会训练时，它仍然让我感到谦卑。

2:27:13 Anyway, I hope that the folks who occasionally get swept up in the chanting online crowds that want to tear down others don’t lose themselves in it too much.

无论如何，我希望那些偶尔被网络上高喊口号、想要拆毁他人的人所卷入的人们不要在其中太迷失自己。

In the end, I still think there’s more good than bad in people.

归根结底，我仍然认为人性中的善多于恶。

But we’re all each of us a mixed bag.

但我们每个人都是一个混合体。

I know I am very much flawed.

我知道我有很大的缺陷。

I speak awkwardly.

我说话很尴尬。

I sometimes say stupid shit.

我有时会说些蠢话。

I can get irrationally emotional.

我可能会变得非理性的情绪化。

I can be too much of a dick when I should be kind.

当我应该友善时，我却变得太混蛋了。

I can lose myself in a biased rabbit hole before I wake up to the bigger, more accurate picture of reality.

在我醒来看到更大、更准确的现实图景之前，我可能会迷失在一个有偏见的兔子洞里。

I’m human and so are you for better or for worse, and I do still believe we’re in this whole beautiful mess together.

我是人类，你也是人类，无论好坏，我仍然相信我们在一起处于这一切美丽的混乱之中。

I love you all.

我爱你们所有人。