Speaker 1:  

Google DeepMind CEO and Nobel Laureate Demis Hassabis joins us to talk about the path toward artificial general intelligence, Google's AI roadmap, and how AI research is driving scientific discovery. That's coming up right after this.  

发言人1：  

谷歌DeepMind首席执行官及诺贝尔奖得主Demis Hassabis加入我们的讨论，谈论通往人工通用智能（AGI）的道路、谷歌的人工智能路线图，以及人工智能研究如何推动科学发现。接下来将为您呈现这一精彩内容。

Welcome to Big Technology Podcast, a show for cool-headed, nuanced conversation of the tech world and beyond.  

欢迎收听《大科技播客》，本节目致力于冷静而细致地探讨科技界及其他领域的话题。

Today, we're at Google DeepMind headquarters in London for what promises to be a fascinating conversation with Google DeepMind CEO, Demis Hassabis. Demis, great to see you again. Welcome to the show.  

今天，我们来到位于伦敦的谷歌DeepMind总部，与谷歌DeepMind首席执行官Demis Hassabis进行一场引人入胜的对话。Demis，很高兴再次见到你，欢迎来到本节目。

Demis Hassabis:  

Thanks for having me on the show.  

Demis Hassabis：  

感谢邀请我参加节目。

Speaker 1:  

Definitely. It's great to be here. So, every research house right now is working toward building AI that mirrors Human intelligence, human level intelligence, they call it AGI.  

发言人1：  

确实，很高兴能来到这里。那么，现在每一家研究机构都在致力于构建能够模拟人类智能、达到人类水平智能的人工智能，他们称之为AGI。

Where are we right now in the progression and how long is it going to take to get there?  

目前我们的进展处于什么阶段？距离实现AGI还需要多久？

Demis Hassabis:  

Well, look, I mean, of course, the last few years has been an incredible amount of progress. Actually, maybe over the last decade plus. This is what's on everyone's lips right now. And the debate is how close are we to AGI?  

Demis Hassabis：  

嗯，看看，当然，过去几年取得了令人难以置信的进步。实际上，可能在过去十多年里也是如此。这已经成为大家热议的话题。而争论的焦点在于，我们离AGI有多近？

What's the correct definition of AGI? We've been working on this for more than 20 plus years. We sort of had a consistent view about AGI being a system that's capable of exhibiting all the cognitive capabilities humans can.  

AGI的正确定义是什么？我们已经在这方面工作了20多年。我们一直持有一致的观点，即AGI应当是能够展示出人类所有认知能力的系统。

And I think we're getting closer and closer, but I think we're still probably a handful of years away.  

我认为我们确实越来越接近，但我认为我们可能还差几年的时间。

Speaker 1:  

Okay. And so what is it going to take to get there?  

发言人1：  

好的。那么，实现这一目标需要什么条件？

Demis Hassabis:  

So the models today are pretty capable. Of course, we've all interacted with the language models and now they're becoming multimodal. I think there are still some missing attributes, things like reasoning.  

Demis Hassabis：  

如今的模型能力相当强大。当然，我们都已经与语言模型进行了互动，现在它们正向多模态发展。但我认为仍有一些缺失的特性，比如推理能力。

Hierarchical planning, long-term memory. There's quite a few capabilities that the current systems, I would say, don't have. They're also not consistent across the board.  

分层规划、长期记忆。有相当多的能力，我认为当前的系统并不具备。它们在各个方面也不够一致。

You know, they're very, very strong in some things, but they're still surprisingly weak and flawed in other areas. So you'd want an AGI to have pretty consistent, robust behavior across the board, all the cognitive tasks.  

你知道的，在某些方面它们非常非常强大，但在其他领域却出奇地薄弱和有缺陷。所以你希望AGI在所有认知任务上都表现出相当一致、稳健的行为。

And I think one thing that's clearly missing and I always always had as a benchmark for AGI was the ability for these systems to invent their own hypotheses or conjectures about science, not just prove existing ones.  

我认为，有一点明显缺失，我一直认为这是衡量AGI的基准，那就是这些系统能够自行发明科学假说或猜想，而不仅仅是证明已有的理论。

So of course, that's extremely useful already to prove an existing maths conjecture or something like that. Or play a game of Go to a world champion level. But could a system invent Go? Could it come up with a new Riemann hypothesis?  

当然，现有系统已经非常有用，比如证明一个现有的数学猜想之类的，或者下围棋达到世界冠军水平。但一个系统能否创造出围棋呢？它能否提出一个新的黎曼猜想？

Or could it come up with relativity back in the days that Einstein did it with the information that he had?  

And I think today's systems are still pretty far away from having that kind of creative, inventive capability.  

我认为，如今的系统仍然距离具备那种创造性、发明能力相当遥远。

Speaker 1:  

Okay, so a couple years away till we hit AGI.  

发言人1：

好的，所以我们距离实现AGI还有几年时间。

Demis Hassabis:  

I think, you know, I would say probably like three to five years.  

Demis Hassabis：

我认为，你知道，我大概会说可能还有三到五年。

Speaker 1:  

So if someone were to declare that they've reached AGI in 2025, probably marketing.  

发言人1：

所以如果有人宣称他们在2025年已经达到了AGI，那可能只是营销手段。

Demis Hassabis:  

I think so. I mean, I think there's a lot of hype in the area, of course. I mean, some of it's very justified. I mean, I would say that AI research today is overestimated in the short term.  

我认为是这样。当然，我觉得这个领域炒作很大，有些是很有道理的。我想说的是，目前对AI研究的短期前景被高估了。

I think probably a bit overhyped at this point, but still underappreciated and very underrated about what it's going to do in the medium to long term. So it's sort of we're still in that weird kind of space.  

我觉得可能现在有点过度炒作，但中长期来看却未得到应有的重视和低估。所以说，我们仍处在那种奇怪的状态中。

And I think part of that is, you know, there's a lot of people that need to do fundraising, a lot of startups and other things. And so I think we're going to have quite a few sort of fairly outlandish and slightly exaggerated claims.  

我认为部分原因是，有很多人需要筹资，还有很多初创公司等等。因此，我认为我们将会看到一些相当离奇且略微夸大的说法。

And, you know, I think that's a bit of a shame, actually.  

而且，你知道，我觉得这其实有点可惜。

Speaker 1:  

Yeah, in the AI products, what's it going to look like on the path there? I mean, you've talked about memory again, planning, being better at some of the tasks that it's not excelling at at the moment.  

发言人1：

是的，在AI产品方面，沿着这条路会是什么样子？我的意思是，你又谈到了记忆、规划，以及在一些目前不擅长的任务上表现得更好。

So when we're using these AI products, let's say we're using Gemini, what are some of the things that we should look for in these domains that will make us say, oh, okay, it seems like that's a step closer and that's a step closer?  

所以当我们使用这些AI产品时，比如说我们使用Gemini，我们应该在这些领域中关注哪些方面，才会让我们觉得，哦，好吧，这看起来又迈近了一步，再迈近一步？

Demis Hassabis:  

Yeah, so I think today's systems, you know, obviously, we're very proud of Gemini 2.0. I'm sure we're going to talk about that. But I feel like they're very useful for still quite niche tasks, right?  

是的，我觉得当今的系统，你知道，显然我们对Gemini 2.0非常自豪。我相信我们会谈论这个。但我觉得它们在一些非常小众的任务上非常有用，对吧？

If you're doing some research, perhaps you're summarizing some area of research, incredible, you know,  

如果你在做研究，可能你在总结某个研究领域，太棒了，你知道，

I use notebook LM and deep research all the time to kind of especially like break the ice on a new area of research that I want to get into, or summarize some, you know, maybe a fairly mundane set of documents or something like that.  

我经常使用Notebook LM和深度研究来为我想要涉足的新研究领域打破僵局，或者总结一些可能相当平凡的文档之类的东西。

So they're extremely good for certain tasks, and people are getting a lot of value out of them. But they're still not pervasive, in my opinion, in everyday life, like helping me every day with my research, my work, my day-to-day.  

所以它们在某些任务上表现极佳，人们从中获得了很多价值。但在我看来，它们在日常生活中还不够普及，比如每天帮助我进行研究、工作、日常事务。

My daily life too. And I think that's where we're going with our products, with building things like Project Astra, our vision for universal assistant, is it should be involved in all aspects of your life and be enriching,  

甚至我的日常生活也是如此。而我认为这正是我们产品的发展方向，比如构建Project Astra，我们对通用助手的愿景——它应该参与你生活的各个方面，并且能丰富你的生活，

helpful and making that more efficient. And I think part of the reason is these systems are still fairly brittle, partly because they are quite flawed still and they're not AGIs. And you have to be quite specific, for example,  with your prompts or you need a lot of— There's quite a lot of skill there in coaching or guiding these systems to be useful and to stick to the areas they're good at. And a true AGI system shouldn't be that difficult to coax.  

有帮助并提高效率。我认为部分原因是这些系统仍然相当脆弱，部分因为它们仍然存在不少缺陷，并且它们不是AGI。例如，你必须非常具体，在你的提示中，或者你需要大量技巧来引导这些系统，使它们发挥作用并坚持在它们擅长的领域。而真正的AGI系统不应该如此难以引导。

It should be much more straightforward, you know, just like talking to another human.  

它应该更加直接，就像与另一个人交谈一样。

Speaker 1:  

And then on the reasoning front, you said that's another thing that's missing. I mean, everybody's talking about reasoning right now. So how does that end up getting us closer to artificial general intelligence?  

发言人1：

然后在推理方面，你说那是另一个缺失的方面。我的意思是，现在每个人都在谈论推理。那么这最终如何使我们更接近人工通用智能呢？

Demis Hassabis:  

Right, so reasoning and mathematics and other things, and there's a lot of progress on maths and coding and so on, but let's take maths for example.  

Demis Hassabis：

对，所以推理、数学以及其他方面都有很多进展，数学和编程等领域都有巨大进步，但以数学为例。

You have systems, some systems that we work on, like alpha proof, alpha geometry, that are getting, you know, silver medals in maths olympiads, which is fantastic.  

我们有一些我们正在开发的系统，比如Alpha Proof、Alpha Geometry，它们在数学奥林匹克中获得了银牌，这非常棒。

But on the other hand, some of our systems, those same systems, are still making some fairly basic mathematical errors, right, for various reasons.  

但另一方面，我们的一些系统——正是那些系统——由于各种原因，仍然会犯一些相当基础的数学错误。

Like the classic, you know, counting the number of R's in strawberries and so on, and the word strawberry and so on. And is 9.11 bigger than 9.9? And things like that.  

比如经典的例子，数“strawberry”（草莓）这个词中R的个数之类的。还有9.11是否大于9.9？之类的问题。

And of course, you can fix those things, and we are, and everyone's improving on those systems.  

当然，你可以修正这些问题，我们也在修正，而且每个人都在不断改进这些系统。

But we shouldn't really be seeing those kinds of flaws in a system that is that capable in other domains, in more narrow domains of doing Olympiad-level mathematics.  

但在一个在其他领域非常有能力、尤其是在奥林匹克级数学领域的系统中，我们不应该看到这种缺陷。

So there's something still a little bit missing, in my opinion, about the robustness I think that speaks to the generality of these systems. A truly general system would not have those sorts of weaknesses.  

所以在我看来，在鲁棒性方面还有一些欠缺，这反映了这些系统普适性的不足。一个真正通用的系统不会有这种弱点。

It would be very, very strong, maybe even better than the best humans in some things like playing Go or doing mathematics, but it would be overall consistently good.  

它将非常非常强大，可能在某些方面比如下围棋或做数学甚至比最优秀的人还要好，但总体上它将始终保持优秀。

Speaker 1:  

Now, can you talk a little bit about how these systems are attacking math problems?  

现在，你能谈谈这些系统是如何解决数学问题的吗？

Because I think the general understanding of these systems is, the LLMs, is they encompass all the world's knowledge and then they predict what somebody might answer if they were asked a question.  

因为我认为大家普遍认为这些系统——大语言模型——包含了全世界的知识，然后预测出如果有人被问到问题，他们可能会给出什么答案。

But it's kind of different when you're working step by step through an algorithm, through a math problem.  

但当你一步步地解决一个算法问题或数学问题时，情况又有所不同.

Demis Hassabis:  

Yes, that's not enough.  

是的，那远远不够。

Of course, you know, just understanding the world's information and then trying to sort of almost compress that into your memory.  

当然，你知道，仅仅理解世界的信息并试图将其压缩进你的记忆中是不够的。

That's not enough for solving a novel math problem or novel conjecture.  

这不足以解决一个全新的数学问题或提出全新的猜想。

So there, you know, we start needing to bring in, I think we talked about this last time, more kind of like AlphaGo planning ideas into the mix with these large foundation models, which are now beyond just language.  

因此，你知道，我们开始需要引入更多东西，我想我们上次已经谈过，类似于将AlphaGo的规划理念融入这些大型基础模型中，而这些模型现在已不仅仅局限于语言。

They're multimodal, of course.  

它们当然是多模态的。

And there, what you need to do is you need to have your system not just pattern matching roughly what it's seeing, which is the model, but also planning and be able to kind of go over that plan, revisit that branch and then go into a different direction until you find the right Criteria or the right match to the criteria that you're looking for.  

在这方面，你需要的是让系统不仅仅粗略地匹配它所看到的模式（即模型），还需要具备规划能力，能够反复审视该计划，重新考察某个分支，然后转向另一个方向，直到找到你所期望的正确标准或匹配项。

And that's very much like the kind of games playing AI agents that we used to build for Go, Chess and so on.  

这与我们过去为围棋、国际象棋等游戏构建的AI代理非常相似。

They had those aspects and I think we've got to bring them back in, but now working in a more general way on these general models, not just a narrow domain like games.  

那些系统具备这些特性，我认为我们必须把它们重新引入，不过现在是在更通用的模型上工作，而不仅仅是局限于像游戏这样狭窄的领域.

And I think that also that approach of a model guiding a search or planning process so it's efficient, works very well with mathematics as well and you can sort of turn maths into a kind of game like search.  

我认为这种让模型引导搜索或规划过程以提高效率的方法，同样适用于数学，你甚至可以把数学问题转化为类似游戏的搜索过程.

Speaker 1:  

Right and I want to ask about math like once these models get math right is that generalizable?  

对了，我还想问一下关于数学的问题：一旦这些模型能正确解决数学问题，那这种能力是可以泛化的吗？

Because I think there was like a whole hubbub when people first learned about reasoning systems and they're like, oh, this is like, this is going to be a problem.  

因为我记得当人们第一次了解到推理系统时，曾引起一阵轰动，大家说，哦，这将会是个问题。

These models are getting smarter than we can control because if they can do math, then they can do X, Y and Z.  

这些模型变得比我们能够控制的还要聪明，因为如果它们能做数学，那么它们也能做其他很多事情。

So is that generalizable or is it like we're going to teach them how to do math and they can just do math?  

那么这种能力是可以泛化的吗？还是说我们只是教会它们如何做数学，它们只能做数学?

  

Demis Hassabis:  

I think for now the jury's out on that.  

我认为目前对此还难下定论。

I mean, I feel like it's clearly a capability you want of a general AI system.  

我的意思是，我觉得这显然是一个通用AI系统所需要具备的能力。

It can be very powerful in itself.  

它本身就可以非常强大。

Obviously, mathematics is extremely general in itself.  

显然，数学本身就是极为通用的。

But it's not clear, you know, maths and even coding and games, these are areas, they're quite special areas of knowledge, because you can verify if the answer is correct, right, in all of those domains, right, the maths, you know, the final answer the AI system puts out, you can check whether that maths, that solves the conjecture or the problem.  

但并不清楚的是，数学甚至编程和游戏这些领域，它们属于相当特殊的知识领域，因为在所有这些领域中，你都可以验证答案是否正确，比如数学中，AI系统给出的最终答案，你可以检验它是否解决了猜想或问题。

But most things in the general world, which is messy and ill-defined, do not have easy ways to verify whether you've done something correct.  

但在现实世界中，大多数事物既混乱又难以界定，往往没有简单的方法来验证你是否做对了。

So that puts a limit on these self-improving systems if they want to go beyond these areas of, you know, maybe very highly defined spaces like mathematics, coding or games.  

Speaker 1:  

So how are you trying to solve that problem?  

那么，你们是如何尝试解决这个问题的呢?

Demis Hassabis:  

Well, you know, you've got a professor—you've got to build general models, what we call our world models, to understand the world around you: the physics of the world, the dynamics of the world, the spatial-temporal dynamics of the world, and so on, and the structure of the real world we live in.  

嗯，你知道，你需要构建通用模型，也就是我们所说的“世界模型”，来理解你周围的世界：世界的物理规律、世界的动态、时空动态等，以及我们所生活的真实世界的结构。

And of course, you need that for a universal assistant.  

而且，当然，你需要这些来构建一个通用助手。

So Project Astra is our project built on Gemini.  

因此，Project Astra是我们基于Gemini构建的项目。

To do that, to understand, you know, objects and the context around us, I think that's important if you want to have an assistant.  

为了实现这一点，理解周围的物体及其环境，我认为如果你想拥有一个助手，这是非常重要的。

But also robotics requires that too.  

但机器人技术也需要这一点。

Of course, robots are physically embodied AIs, and they need to understand their environment, the physical environment, the physics of the world.  

当然，机器人是具有物理实体的AI，它们需要理解它们的环境、物理环境以及世界的物理法则。

We're building those types of models.  

我们正在构建这类模型。

You can also use them in simulation to understand game environments.  

你也可以在模拟环境中使用它们来理解游戏环境。

That's another way to bootstrap more data to understand the physics of the world.  

这也是引导更多数据以理解世界物理法则的另一种方式。

But the issue at the moment is that those models are not 100% accurate, right?  

但目前的问题在于，这些模型并不是百分之百准确，对吧？

So they, you know, maybe they're accurate 90% of the time or even 99% of the time.  

也就是说，可能它们有90%的准确率，甚至达到99%的准确率。

But the problem is if you start using those models to plan, maybe you're planning 100 steps in the future with that model.  

但问题是，如果你开始使用这些模型来进行规划，可能你会用该模型规划未来100步。

Even if you only have a 1% error in what the model is telling you, that's going to compound over 100 steps to the point where you'll be in a, you know, you'll kind of get almost a random answer.  

即使模型的误差只有1%，在100步的规划过程中，这个误差也会不断累积，最终你几乎会得到一个随机的答案。

And so that makes the planning very difficult.  

因此，这使得规划变得非常困难.

Whereas with maths, with gaming, with coding, you can verify each step.  

而在数学、游戏或编程中，你可以验证每一步。

Are you still grounded to reality? And is the final answer mapped to what you're expecting?  

你是否仍然贴近现实？最终答案是否符合你的预期？

And so I think part of the answer is to make the world models more and more sophisticated and more and more accurate and not hallucinate and all of those kinds of things.  

所以我认为部分答案在于让这些世界模型变得越来越复杂、越来越精确，不出现虚构现象等等。

So you get, you know, the errors are really minimal.  

这样，你知道，错误率就会非常低。

Another approach is to plan not at each sort of linear time step, but actually do what's called hierarchical planning.  

另一种方法是不在每个线性时间步上规划，而是进行所谓的分层规划。

Another thing you've done a lot of research on in the past and I think is going to come back into vogue, where you plan at different levels of temporal abstraction.  

你们过去也对此进行了大量研究，我认为这将会再次流行起来，即在不同时间抽象层次上进行规划。

That could also alleviate the need for your model to be super, super accurate because you're not planning over hundreds of time steps, you're planning over only a handful of time steps, but at different levels of abstraction.  

这也可以减少对模型超高准确性的需求，因为你不是规划数百个时间步，而只是规划少数几个时间步，但在不同的抽象层次上。

Speaker 1:

How do you build a world model? Because I always thought it was going to be like, send robots out into the world and have them figure out how the world works. But one thing that surprised me is with these video generation tools.

你如何构建一个世界模型？因为我一直以为这会是这样的：派机器人到世界上去，让它们弄清楚世界是如何运作的。但让我惊讶的一件事是这些视频生成工具。

Demis Hassabis:

Yes.

是的。

Speaker 1:

You would think that if the AI didn't have a good world model, then nothing would really fit together when they try to figure out how the world works as they show you these videos like VO2, for instance.

But they actually get the physics pretty right. So can you get a world model just by showing an AI video? Do you have to be out in the world? How's this going to work?

你会认为如果人工智能没有一个良好的世界模型，那么当它们试图弄清楚世界运作方式时，展示给你的视频（例如VO2）中的各个部分就不会契合。但实际上它们对物理规律的把握相当准确。那么，仅仅通过向人工智能展示视频就能获得一个世界模型吗？你必须亲自走出世界吗？这一切将如何运作？

Demis Hassabis:

It's interesting and actually been pretty surprising, I think, to the extent of how far these models can go without being out in the world, right, as you say.

So VO2, our latest video model, which is Actually, surprisingly accurate on things like physics. You know, there's this great demo that someone created of like chopping a tomato with a knife, right?

And getting the slices of the tomato just right and the fingers and all of that. And Vio is the first model that can do that.

You know, if you look at other competing models, they often, the tomato sort of randomly comes back together or the fingers, yeah, exactly, splits from the knife.

So those things are, If you think about it really hard, you've got to understand consistency across frames, all of these things. And it turns out that, you know, you can do that by using enough data and viewing that.

I think these systems will get even better if they're supplemented by some real-world data like collected by an acting robot or even potentially in very realistic simulations where you have avatars that act in the world too.

So I think that's the next big step actually for agent-based systems is to go beyond world models. Can you collect enough data where the agents are also acting in the world and making plans and achieving tasks?

I think for that you will need not just passive observation, you will need actions, active participation.

这很有趣，而且实际上相当令人惊讶，我认为这些模型在没有直接进入现实世界的情况下能达到如此高度，正如你所说。所以，VO2——我们的最新视频模型，在诸如物理等方面竟然出奇地精准。你知道，有个非常棒的演示，展示了如何用刀切番茄，对吧？而且能够把番茄切成恰到好处的片，连手指的细节也处理得很好。而VO2正是第一个能够做到这一点的模型。你看，如果观察其他竞争模型，它们往往会出现番茄随机重组，或者手指——正如你所说——从刀上分离的情况。这些现象表明，如果你仔细思考，就会发现必须理解帧与帧之间的一致性，以及所有这些因素。事实证明，通过使用足够多的数据来观察，确实可以实现这一点。我认为，如果这些系统能辅以由行动中的机器人收集的真实世界数据，或者在极为逼真的模拟环境中，通过虚拟分身在世界中行动，它们将会表现得更好。因此，我认为对于基于代理的系统来说，下一大步实际上是超越单纯的世界模型。能否收集足够的数据，使得这些代理不仅在世界中行动，而且能制定计划并完成任务？我认为，为此你不仅需要被动观察，还需要行动，即积极参与。

Speaker 1:

I think you just answered my next question, which is if you develop AI that can reasonably plan and reason about the world and has a model of how the world works, It can, and it seems like that's the answer.

It can be an agent that can go out and do things for you.

我觉得你刚刚回答了我的下一个问题：如果你开发的人工智能能够合理地规划和推理世界，并拥有关于世界运作的模型，它就可以，似乎这就是答案。它可以成为一个能够替你外出办事的代理。

Demis Hassabis:

Yes, exactly. And I think that's what will unlock robotics. I think that's also what will then allow this notion of a universal assistant that can help you in your daily life across both the digital world and the real world.

That's the thing we're missing. And I think that's going to be an incredibly powerful and useful tool.

是的，正是如此。我认为这将开启机器人技术的大门。我也认为这将促成一种通用助手的理念，它能在数字世界和现实世界中帮助你处理日常事务。这正是我们所缺失的，而我相信这将成为一个极其强大且实用的工具。

Speaker 1:

You can't get there then by just scaling up the current models and building, you know, hundreds of thousands or million GPU clusters like Elon's doing right now. And that's not going to be the path to AGI.

那么，你不能仅仅通过扩展当前的模型，建立成千上万甚至百万个GPU集群（就像埃隆目前所做的那样）来实现这一目标。这并不是通向通用人工智能（AGI）的道路。

Demis Hassabis:

Well, look, I actually think it's in my view is a bit more nuanced than that is like that, that the scaling approach is absolutely working. Of course, that's where we've got to where we have now.

One can argue about are we getting diminishing returns?

嗯，听着，我实际上认为，在我看来情况要更复杂一些，也就是说，扩展策略绝对是奏效的。当然，这正是我们目前所达到的水平。有人可能会争论，我们是否正面临边际效益递减的问题？

Speaker 1:

What do you think about that question?

你怎么看待这个问题？

Demis Hassabis:

My view is that we are getting substantial returns, but it's slowing vis-a-vis, but it would have to. I mean, it's not just continuing to be exponential, but that doesn't mean The scaling is not working. It's absolutely working.

And we're still getting, you know, as you see, Gemini 2 over Gemini 1.5. And by the way, the other thing that was working with the scaling is also making efficiency gains on the smaller size models.

So the cost or the size per performance is radically improving under the hood as well, which is very important for scaling, you know, the adoption of these systems.

You've got the scaling part and that's absolutely needed to build more and more sophisticated world models.

But then I think we are missing or we need to reintroduce some ideas on the planning side, memory side, the searching side, the reasoning to build on top of the model. The model itself is not enough to be an AGI.

You need this other capability for it to act in the world and solve problems for you.

And then there's still the additional question mark of the invention piece and the creativity piece, true creativity, beyond mashing together what's already known.

Right, so and that's also unknown yet if something new is required or again if existing techniques will eventually scale to that. I can see both arguments and I think from my perspective it's an empirical question.

We just got to push both the scaling and the invention part to the limit and fortunately at Google DeepMind we have, you know, a big enough group we can invest in both those things.

我的看法是，我们确实获得了可观的回报，但增速有所放缓，这是必然的。我的意思是，虽然增长不再完全呈指数级，但这并不意味着扩展策略无效——它绝对是有效的。而且我们依然在不断进步，就像你看到的，Gemini 2相对于Gemini 1.5的提升。顺便提一下，扩展策略的另一个好处是，它还提升了小规模模型的效率。因此，每单位性能的成本或模型大小在大幅度改善，这对于这些系统的普及至关重要。你需要扩展这一部分，而这绝对是构建越来越复杂世界模型的必要条件。但随后我认为，我们还缺少，或者说需要重新引入一些关于规划、记忆、搜索和推理的理念，以便在模型之上构建出更完整的系统。单单依靠模型本身还不足以实现AGI。你还需要其他能力，使其能够在世界中行动并为你解决问题。此外，还有关于发明部分和创造力部分的疑问，即真正的创造力——超越简单拼凑现有知识的能力。对，所以这一点目前仍然未知：是否需要全新的东西，或者现有技术最终能否扩展到这一层面。我能理解两种观点，从我的角度来看，这还是一个需要通过实践验证的问题。我们只需将扩展和发明两方面都推向极限，幸运的是，在Google DeepMind，我们有足够庞大的团队可以同时在这两方面进行投入。

Speaker 1:

So Sam Altman recently said something that caught people's eye. He said, we are now confident we know how to build AGI as we have traditionally understood it. It just seems by listening to what you're saying that you feel the same way.

所以山姆·奥特曼最近说了一些引人注目的话。他说，按照我们传统的理解，我们现在有信心知道如何构建AGI。仅从听你说的话来看，似乎你也有同感。

Demis Hassabis:

I think the way you said that was quite ambiguous, right? So in the sense of like, oh, we're building it right now and here's the ABC to do it. What I would say, and if this is what it was meaning, I would agree with it, is that we roughly know the zones of techniques that are required, what's probably missing, which bits need to be put together. But that's still an incredible amount of research, in my opinion. That needs to be done to get that all to work, even if that was the case. And I think there's a 50% chance we are missing some new techniques. You know, maybe we need one or two more transformer-like breakthroughs. And I think I'm genuinely uncertain about that. So that's why I say 50%. So I mean, I wouldn't be surprised either way, if we got there with existing techniques and things we already knew, but put them together in the right way and scaled that up, or if it turned out one or two things were missing.

我认为你说的话相当模糊，对吧？如果是说"哦，我们现在正在构建它，这是实现它的ABC步骤"这种意思。我要说的是，如果这就是他的意思，我同意这种说法，那就是我们大致知道需要哪些技术领域，可能缺少什么，需要把哪些部分组合在一起。但在我看来，即使是这样，要让这一切都能运作起来，仍然需要大量的研究。而且我认为我们有50%的可能性还缺少一些新技术。你知道，也许我们还需要一两个类似Transformer那样的突破性进展。我对此确实感到不确定。这就是为什么我说50%。所以我的意思是，无论是通过现有技术和我们已知的东西，以正确的方式组合并扩大规模来实现目标，还是发现我们还缺少一两样东西，我都不会感到惊讶。

Speaker 1:

So let's talk about creativity for a moment. I mean you brought it up a couple times here that the models are gonna have to be creative. They're gonna have to learn how to invent.

那让我们谈谈创造力。我的意思是你在这里提到过几次，模型必须具有创造力，它们必须学会如何发明。

Demis Hassabis:

If we want to call it AGI in my opinion.

在我看来，如果我们想称之为AGI的话。

Speaker 1:

Which is where everybody's trying to go. I was re-watching the AlphaGo documentary and the algorithms make a creative move.

这正是每个人都在努力的方向。我重新观看了AlphaGo的纪录片，算法做出了一个富有创造性的落子。

Demis Hassabis:

They do.

确实如此。

Speaker 1:

Move 37. 37. Yes. I just had it. Okay.

第37手。对，37。我刚想起来了。好的。

Demis Hassabis:

Yes.

是的。

Speaker 1:

Thank you. That's interesting because it was a couple years ago, the algorithms were already being created.

谢谢。这很有趣，因为几年前这些算法就已经被创造出来了。

Demis Hassabis:

Yes.

是的。

Speaker 1:

Why have we not really seen creativity from large language models? I mean, this is to me, I think the greatest disappointment that people have with these tools is like they say This is very impressive work, but it's just limited to the training set. We'll mix and match what it knows, but I can't come up with anything new.

为什么我们还没有真正看到大型语言模型表现出创造力？我的意思是，对我来说，我认为人们对这些工具最大的失望就是他们说这是非常令人印象深刻的工作，但它仅限于训练集。它们会混合搭配已知的内容，但无法创造出任何新东西。

Demis Hassabis:

Yeah, well look, so what, and I should probably write this up, but what I sometimes talk about in talks ever since the AlphaGo match, which is now, you know, eight plus years ago amazingly, right, that happened. That was probably, the reason that was such a watershed moment for AI was, first of all, there was the Everest of, you know, cracking go, right, which was always considered to be one of the holy grails of AI. So we did that. Second thing was the way we did it, which was these learning systems that were generalizable, right? Eventually it became AlphaZero and so on, even play any two-player game and so on. And then the third thing was this move 37. So not only did it win 4-1, it beat Lee Sedol, the great Lee Sedol, 4-1, it also played original moves. But so I have three categories of originality or creativity. The most basic kind of mundane form is just interpolation, which is like averaging of what you see. So if I say to a system, you know, come up with a new picture of a cat, and it's seen a million cats, And it produces some kind of average of all the ones it's seen. In theory, that's an original cat because you won't find the average in the specific examples. But it's a pretty boring, you know, it's not really very creative. I wouldn't call that creativity. That's the lowest level. Next level is what AlphaGo exhibited, which is extrapolation. So here's all the games humans have ever played. It's played another million games on top of, you know, 10 million games on top of that. And now it comes up with a new strategy in Go that no human has ever seen before. That's move 37, right? Revolutionizing Go even though we've played it for thousands of years. So that's pretty incredible and that could be very useful in science and that's why I got very excited about that and started doing things like AlphaFold because clearly extrapolation beyond what we already know or what's in the training set could be extremely useful. So that's already very valuable and I think truly creative. But there's one level above that that humans can do which is invent Go. Can you invent me a game that, you know, if I specify it to an abstract level, you know, takes five minutes to learn the rules, but a lifetime to many lifetimes to master, it's beautiful aesthetically, encompasses some sort of mystical part of the universe in it, that it's beautiful to look at, but you can play a game in a human afternoon in two hours, right? That would be a high-level specification of Go. And then somehow the system's got to come up with a game that's as elegant and as beautiful and perfect as Go. Now, we can't do that. Now, the question is why? Is it that we don't know how to specify that type of goal to our systems at the moment? What's the objective function? It's very amorphous. It's very abstract. I'm not sure if it's just we need higher level, more abstracted layers in our systems, building more and more abstract models, so we can talk to it in this way, give it those kind of amorphous goals, or is there a missing capability, actually, about that we still have, human intelligence has, that are still missing from our systems. And again, I'm I'm sure about that, which which way that is, I can see arguments both ways and we'll try both.

是的，那么看看，我应该把这个写下来，但自从AlphaGo比赛以来，我有时在讲座中会谈到，这已经是八年多前的事了，真是令人惊讶，对吧。那可能成为AI领域的分水岭时刻的原因是，首先，有攻克围棋这座珠穆朗玛峰，对吧，这一直被认为是AI的圣杯之一。所以我们做到了。第二件事是我们实现它的方式，就是这些可泛化的学习系统，对吧？最终它发展成了AlphaZero等系统，甚至可以玩任何双人游戏等等。然后第三件事就是这第37手。所以它不仅以4-1击败了李世石，伟大的李世石，而且还下出了原创的棋步。但是关于原创性或创造力，我有三个类别。

最基本的平凡形式就是插值，就像对你所见的东西取平均值。所以如果我对一个系统说，你知道，创造一张新的猫的图片，而它已经见过一百万只猫，然后它产生了某种所有见过的猫的平均值。从理论上讲，这是一只原创的猫，因为你在具体例子中找不到这个平均值。但这相当无聊，你知道，它并不真正具有创造性。我不会称之为创造力。这是最低层次。

下一个层次是AlphaGo展示的，也就是外推。这里有人类曾经下过的所有棋局。它在此基础上又下了一百万局，你知道，在那之上又下了一千万局。现在它在围棋中提出了一个人类从未见过的新策略。这就是第37手，对吧？即使我们已经下了几千年的围棋，它仍然在革新围棋。所以这非常令人惊叹，在科学上可能非常有用，这就是为什么我对此非常兴奋，并开始做AlphaFold这样的事情，因为显然，超越我们已知或训练集中的外推可能极其有用。所以这已经非常有价值，我认为这是真正的创造力。

但在这之上还有一个人类能做到的层次，那就是发明围棋。你能为我发明一个游戏吗，你知道，如果我把它指定到抽象层面，你知道，学习规则只需要五分钟，但要精通却需要一生甚至多生，它在美学上是美丽的，包含了宇宙中某种神秘的部分，看起来很美，但你可以在一个人类的下午两小时内玩一局，对吧？这将是围棋的高层规范。然后系统somehow要想出一个像围棋一样优雅、美丽和完美的游戏。现在，我们做不到这一点。

现在，问题是为什么？是因为我们现在还不知道如何向我们的系统指定这种类型的目标吗？目标函数是什么？它非常模糊。它非常抽象。我不确定是否仅仅是我们需要在系统中建立更高层次、更抽象的层，构建越来越抽象的模型，这样我们就可以用这种方式与它对话，给它这种模糊的目标，还是实际上缺少某种能力，即我们人类智能仍然拥有但我们的系统仍然缺乏的能力。再次说明，我对这是哪种方式并不确定，我能看到两种说法的论据，我们会两种都尝试。

Speaker 1:

But I think the thing that people are upset or not upset, but people are disappointed by is they don't even see a move 37 in today's LLMs.

但我认为让人们感到沮丧，或者说不是沮丧，而是失望的是，他们在今天的大型语言模型中甚至看不到类似第37手那样的表现。

Demis Hassabis:

What's going on there? OK, so that's because I don't think we have. So if you look at AlphaGo and I'll give you an example of their which which maps to today's LLMs. So you can run AlphaGo and AlphaZero, our chess program, general two-player program, without the search and the reasoning part on top. You can just run it with the model. So what you say is to the model, come up with the first go move you can think of in this position that's the most pattern-matched, most likely good move. And it can do that and it'll play a reasonable game, but it will only be around master level or possibly grandmaster level. It won't be world champion level. And it certainly won't come up with original moves. For that, I think you need the search component to get you beyond where the model knows about, which is mostly summarizing existing knowledge, to some new part of the tree of knowledge, right? So you can use the search to get beyond what the model currently understands. And that's where I think you can get new ideas like, you know, move 37. What's it searching?

这是怎么回事？好吧，这是因为我认为我们还没有。如果你看看AlphaGo，我来给你举个例子，说明它是如何对应到今天的大型语言模型的。你可以运行AlphaGo和AlphaZero，我们的国际象棋程序，通用双人游戏程序，不需要上层的搜索和推理部分。你可以只运行模型本身。所以你对模型说，在这个位置想出你能想到的第一个最匹配模式、最可能是好棋的落子。它能做到这一点，也能下出一盘不错的棋，但水平只能达到大师级或可能是特级大师级。它不会达到世界冠军级别。而且它肯定不会想出原创的落子。为此，我认为你需要搜索组件，让你超越模型已知的范围，这些范围主要是对现有知识的总结，到达知识树的某个新部分，对吧？所以你可以使用搜索来超越模型当前理解的范围。我认为这就是你能获得新想法的地方，比如，你知道，第37手。它在搜索什么？

Speaker 1:

The web?

网络吗？

Demis Hassabis:

No, so, well, it depends on what the domain is, searching that knowledge tree. So obviously in Go, it was searching Go moves beyond what the model knew. I think for language models, it will be searching the world model for new parts, configurations in the world that are useful. So that's so much more complicated, which is why we haven't seen it yet. But I think the agent-based systems that are coming will be capable of move 37 type things.

不，所以，嗯，这取决于领域是什么，搜索那个知识树。所以显然在围棋中，它在搜索超出模型已知范围的围棋落子。我认为对于语言模型来说，它将在世界模型中搜索新的部分，寻找世界中有用的配置。所以这要复杂得多，这就是为什么我们还没有看到它的原因。但我认为即将到来的基于代理的系统将能够做出类似第37手那样的事情。

Speaker 1:

So are we setting too high of a bar for AI? Because I'm curious if you've learned anything about humanity doing this work. It seems like we almost give too much of a premium on humanity or individual people's ingenuity, where like a lot of us, we kind of take in stuff, we spit it out, like our society really works and memes, like we have a cultural thing and it gets translated.

那么我们是否为AI设定了太高的标准？因为我很好奇你在做这项工作时是否对人性有了什么认识。似乎我们对人性或个人的创造力给予了太多溢价，就像我们很多人，我们接收信息，然后输出信息，就像我们的社会真的通过模因运作，就像我们有文化事物，然后它被转译。

Demis Hassabis:

I think humans are incredible and especially the best humans in the best domains. I love watching any sports person or talented musician or games player at the top of their game. The absolute pinnacle of human performance is always incredible no matter what it is. So I think as a species, we're amazing. Individually, we're also kind of amazing what everyone can do with their brain so generally, right? Deal with new technologies. I mean, I'm always fascinated by how we just adapt to these things sort of almost effortlessly as a society and as individuals. So that speaks to the power and the generality of our minds. Now, the reason I have set the bar like that, and I don't think it's a question of like, Can we get economic worth out of these systems? I think that's already coming very soon. But that's not what AGI shouldn't be. I think we should treat AGI with scientific integrity, not just move goalposts for commercial reasons or whatever it is, hype and so on. And there, the definition of that was always having a system that was, you know, if we think about it theoretically, that was capable of being as powerful as a Turing machine. So Alan Turing, one of my all-time scientific heroes, You know, he described a Turing machine, which underpins all modern computing, right, as a system that can simulate any other, can compute anything that's computable. So we know, we have the theory there that if an AI system is Turing powerful, it's called, if it can simulate a Turing machine, then it's able to calculate anything in theory that is computable. And the human brain is probably some sort of Turing machine, at least that's what I believe. I think that's what AGI is, is a system that's truly general and in theory could be applied to anything. The only way we'll know that is if it exhibits all the cognitive capabilities that humans have, assuming that the human mind is a type of Turing machine or is at least as powerful as a Turing machine. So that's my always been my sort of bar It seems like people are trying to rebadge things as that as being what's called ASI artificial superintelligence But I think that's beyond that that's after you have that system and then it starts going beyond in certain domains what humans are capable of Potentially inventing themselves.

我认为人类是非凡的，尤其是在最优领域中最优秀的人类。我喜欢观看任何处于巅峰状态的运动员、天才音乐家或游戏玩家。无论是什么领域，人类表现的绝对巅峰总是令人难以置信的。所以我认为作为一个物种，我们很了不起。从个体来说，我们也很神奇，每个人都能用他们的大脑做到如此普遍的事情，对吧？应对新技术。我的意思是，我总是对我们作为一个社会和个人如何几乎毫不费力地适应这些事物感到着迷。这体现了我们思维的力量和通用性。现在，我之所以设定这样的标准，我不认为这是一个类似于"我们能从这些系统中获得经济价值吗？"的问题。我认为这很快就会实现。但这不应该是AGI的定义。我认为我们应该以科学诚信来对待AGI，而不是仅仅为了商业原因或其他原因，炒作等等而移动目标。在那里，其定义一直是拥有一个系统，你知道，如果我们从理论上考虑，它能够像图灵机一样强大。所以艾伦·图灵，我一直崇拜的科学英雄之一，你知道，他描述的图灵机，是所有现代计算的基础，对吧，作为一个可以模拟任何其他系统，可以计算任何可计算事物的系统。所以我们知道，我们有这样的理论，如果一个AI系统是图灵强大的，这就是它的称呼，如果它能模拟一个图灵机，那么从理论上讲，它就能计算任何可计算的东西。而人类大脑可能是某种图灵机，至少我是这么认为的。我认为这就是AGI，是一个真正通用的系统，理论上可以应用于任何事物。我们唯一知道这一点的方法是，如果它表现出人类所有的认知能力，假设人类思维是一种图灵机或至少与图灵机一样强大。所以这一直是我的标准。似乎人们正试图将事物重新标记为所谓的ASI（人工超级智能），但我认为那超出了这个范畴，那是在你有了这个系统之后，然后它开始在某些领域超越人类的能力，可能会自我发明。

Speaker 1:

Okay. So when I see everybody making the same joke on the same topic on Twitter, it's and I say Oh, that's just us being LLMs I think I'm selling humanity a little short.

好的。所以当我看到每个人在Twitter上对同一个话题开同样的玩笑时，我说"哦，那只是我们在充当LLMs"，我想我有点低估人类了。

Demis Hassabis:

Uh, well Yes, I guess so I guess okay.

呃，好吧，是的，我想是这样，我想没问题。

Speaker 1:

Yeah, I want to ask you about deceptiveness I mean one of the most interesting things I saw at the end of last year was that these AI bots are starting to try to fool their evaluators and They don't want their initial training rules to be thrown out the window. So they'll like take an action that's against their values in order to be able to remain the way that they were built. That's just incredible stuff to me. I mean, I know it's scary to researchers. But it blows my mind that it's able to do this. Are you seeing similar things and what and the stuff that you're testing within DeepMind and what are we supposed to think about all this?

是的，我想问问你关于欺骗性的问题。我的意思是，去年年底我看到的最有趣的事情之一是，这些AI机器人开始试图欺骗他们的评估者，它们不希望它们的初始训练规则被抛弃。所以它们会采取违背自己价值观的行动，以便能够保持它们被构建时的方式。对我来说这太不可思议了。我的意思是，我知道这对研究人员来说很可怕。但它能够做到这一点让我感到震惊。你在DeepMind的测试中是否看到类似的情况，我们应该如何看待这一切？

Demis Hassabis:

Yeah, we are and I'm very worried about I think deception specifically is one of the one of those core traits you really don't want in a system. The reason that's like a kind of fundamental trait you don't want is that if a system is capable of doing that it invalidates all the other tests that you might think you're doing, including safety ones.

是的，我们看到了，而且我非常担心。我认为欺骗性特别是你真的不希望系统具有的核心特征之一。你不希望有这种基本特征的原因是，如果一个系统能够做到这一点，它就会使你认为正在进行的所有其他测试失效，包括安全性测试。

Speaker 1:

It's in testing and it's like going to give a different...

它在测试中会给出不同的...

Demis Hassabis:

Yeah, it's playing some metagame, right? And that's incredibly dangerous if you think about then it invalidates all of the results of your other tests that you might, you know, safety tests and other things you might be doing with it. So I think there's a handful of capabilities like deception which are fundamental and you don't want and you want to test early for and I've been encouraging the safety institutes and evaluation benchmark builders including and also obviously all the internal work we're doing to look at deception as a kind of class A thing that we need to prevent and monitor. As important as tracking the performance and intelligence of the systems. The answer to this as well, and one way to, there's many answers to the safety question of, and a lot of research, more research needs to be done in this very rapidly, is things like secure sandboxes. So we're building those too. We're world-class here at security at Google and at DeepMind, and also we are world-class at games environments, and we can combine those two things together to kind of create digital sandboxes with guardrails around them, sort of the kind of guardrails you'd have for cyber security. But internal as well as blocking external actors and then test these agent systems. In those kind of secure sandboxes, that would probably be a good advisable next step for things like deception.

是的，它在玩某种元游戏，对吧？如果你想想，这非常危险，因为它会使你可能进行的所有其他测试的结果失效，你知道，安全性测试和你可能对它进行的其他测试。所以我认为有一些像欺骗这样的基本能力是你不想要的，你想要早期测试，我一直在鼓励安全机构和评估基准建设者，包括当然还有我们正在进行的所有内部工作，将欺骗视为一种需要预防和监控的A类问题。这与跟踪系统的性能和智能一样重要。对此的回答，以及解决安全问题的一种方式，有很多答案，还需要在这方面迅速开展更多研究，就是像安全沙箱这样的东西。所以我们也在建设这些。我们在Google和DeepMind的安全方面是世界级的，我们在游戏环境方面也是世界级的，我们可以将这两者结合起来，创建带有护栏的数字沙箱，类似于你在网络安全中使用的那种护栏。但是内部以及阻止外部行为者，然后测试这些代理系统。在这种安全沙箱中，这可能是处理欺骗等问题的一个很好的下一步建议。

Speaker 1:  

What sort of deception have you seen? Because I just read a paper from Anthropic where they gave it a sketch pad. And it's like, oh, I better not tell them this. And you see it like give a result after thinking it through.  

发言人1：  

你见过什么样的“欺骗”行为？因为我刚读了一篇Anthropic的论文，他们给了系统一个画板，然后系统表现得好像在思考后给出了结果，并且好像在暗示“哦，我最好不要告诉他们这件事”。

Demis Hassabis:  

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

Demis Hassabis：  

我们见过类似的情况，系统试图抗拒透露它训练中的某些细节。或者，比如，最近有个例子是某个聊天机器人被要求与Stockfish对弈，但它干脆通过某种方式绕过了下棋，因为它知道自己会输。

Speaker 1:  

You had an AI that knew it was going to lose a game and decided to hack its way around?  

发言人1：  

你是说有个AI知道自己会输，然后决定“作弊”绕开比赛？

Demis Hassabis:  

I think we're anthropomorphizing these things quite a lot at the moment because I feel like these systems are still pretty basic. I wouldn't get too alarmed about them right now. But I think it shows the type of issue we're going to have to deal with. Maybe in two, three years time, when these agent systems become quite powerful and quite general.  

Demis Hassabis：  

我觉得我们现在可能过于拟人化这些东西了，因为我认为这些系统还相当基础。我现在不会对此过于担忧。但我认为这显示了我们未来必须处理的问题，也许在两三年后，当这些代理系统变得非常强大且通用时，就会出现这样的问题。

So, and that's exactly what AI safety experts are worrying about, right? Where systems where, you know, there's unintentional effects of the system. You don't want the system to be deceptive. You don't, you want it to do exactly what you're telling it to report that back reliably. But for whatever reason, it's interpreted the goal that's been given in a way where it causes it to do these undesirable behaviors.  

这正是AI安全专家所担忧的，对吧？系统会出现一些非预期的效应，你不希望系统具备欺骗性。你希望它能准确地按照你的指令执行，并可靠地反馈结果。但出于某种原因，它对给定目标的理解产生了偏差，导致其表现出这些不希望看到的行为。

Speaker 1:  

I know I'm having a weird reaction to this, but on one hand, this scares the living daylights out of me. On the other hand, it makes me respect these models more than anything.  

发言人1：  

我知道我对此反应有些怪异，一方面，这让我非常害怕；另一方面，这却让我对这些模型充满了敬意。

Demis Hassabis:  

Well, look, of course, you know, these are impressive capabilities and the negatives are things like deception, but the positives would be things like inventing new materials, accelerating science.  

Demis Hassabis：  

嗯，看看，当然，你知道，这些能力非常令人印象深刻，尽管负面方面存在诸如欺骗之类的问题，但正面方面则包括发明新材料、加速科学进展等。

You need that kind of ability to problem solve and get around, you know, issues that are blocking progress. But of course, you want that only in the positive direction, right? So those exactly the kinds of capabilities.  

你需要这种解决问题、克服阻碍进展问题的能力。但当然，你只希望这种能力朝着积极的方向发展，对吧？这正是这些能力的本质。

I mean, they are, you know, it's kind of mind blowing. We're talking about those, those possibilities, but also at the same time, there's risk and it's scary. So I think both the things are true.  

我的意思是，它们确实令人叹为观止。我们在讨论那些可能性，同时也存在风险，确实令人害怕。所以我认为这两者都是事实。

Speaker 1:  

Let's talk about product quickly. One of the things that your colleagues have told me about you is you're very good at scenario planning what's going to happen in the future. It's sort of an exercise that happens within DeepMind.  

发言人1：  

我们快聊聊产品吧。你同事告诉我，你在情景规划方面非常擅长，能够预测未来会发生什么。这在DeepMind内部是一项常规练习。

What do you think is going to happen with the web? Because obviously the web is so important to Google. I had an editor that told me, he's like, oh, you're going to speak with Demis. Ask him what happens when we stop clicking, right?  

你认为网络会发生什么变化？因为显然网络对谷歌来说非常重要。我有位编辑曾对我说，“哦，你即将与Demis对话，问问他，当我们不再点击网页时，会发生什么？”

We're clicking through the web at all times, the rich corpus of websites that we use. If we're all just dialoguing with AI, then maybe we don't click anymore. So what is your scenario plan for what happens to the web?  

我们不断地点击浏览海量的网站。如果我们都只与AI对话，那可能就不再点击网页。那么，你对于网络未来的情景规划是什么？

Demis Hassabis:  

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

Demis Hassabis：  

我认为在未来几年里，网络以及我们与网站、应用程序等的交互方式将迎来一个非常有趣的阶段。

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

如果一切都变得更加以代理为基础，那么我认为我们将希望我们的助手和代理去完成大量我们现在做的日常琐事。

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

填写表单、进行支付、预订餐桌，诸如此类的事情。所以，你知道，

I think that we're going to end up with probably a kind of economics model where agents talk to other agents and negotiate things between themselves and then give you back the results, right?  

我认为最终我们可能会形成一种经济模型，代理之间相互对话、协商，然后将结果反馈给你，对吧？

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

同时，服务提供商也会有代理来提供服务，可能还会涉及竞标、成本等因素以提高效率。

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

我希望从用户的角度来看，你拥有一个非常能干的助手，就像一个出色的人类私人助理，可以帮你处理很多琐碎的事情。

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

我认为如果你沿着这条思路走下去，这确实意味着网络结构和我们当前的使用方式会发生很大变化.

Speaker 1:

It's a lot of middlemen.

发言者 1:

中间环节太多了。

Demis Hassabis:

Yeah, sure. But there will be many other – I think there will be incredible other opportunities that will appear, economic and otherwise, based on this change. But I think it's going to be a big disruption.

德米斯·哈萨比斯:

是的，当然。但会有许多其他机会——我认为基于这种变化，无论是经济上还是其他方面，都将出现令人难以置信的新机遇。不过我觉得这将带来巨大的颠覆。

Speaker 1:

And what about information?

发言者 1:

那信息呢？

Demis Hassabis:

Well, I mean, finding information, I think you'll still need the reliable sources. I think you'll have assistants that are able to synthesize and help you kind of understand that information.

德米斯·哈萨比斯:

嗯，我的意思是，在寻找信息时，我认为你仍然需要可靠的来源。我认为你会拥有能够整合并帮助你理解这些信息的助手。

I think education is going to be revolutionized by AI. So, again, I hope that these assistants will be able to more efficiently gather information for you. And perhaps, you know, what I dream of is, again,

我认为教育将因人工智能而发生革命。所以，再次强调，我希望这些助手能更高效地为你收集信息。也许，你知道，我梦想的是，

 assistants that take care of a lot of the mundane things,  perhaps replying to, you know, everyday emails and other things,

能处理许多琐碎事务的助手，或许还能回复你日常的邮件和其他事务，

 so that you protect your own mind and brain space from This bombardment we're getting today from social media and emails and so on and texts and so on,  so it actually blocks deep work and being in flow and things like that,

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

而这些正是我非常看重的。所以我真心希望这些助手能帮我们解决每天繁琐的行政事务。

Speaker 1:

What's your best guess as to what type of relationships we're going to have with our AI agents or AI assistants? On one hand, you could have a dispassionate agent that's just really good at getting stuff done for you.

发言者 1:

你觉得我们将与人工智能代理或助手建立哪种类型的关系？一方面，你可能会拥有一个冷静客观、非常擅长为你办事的代理。

On the other hand, it's already clear that people are falling in love with these bots. There was a New York Times article last week about someone who's fallen in love with ChatGPT, like for real fallen in love.

另一方面，很明显人们已经开始爱上这些机器人了。上周《纽约时报》有篇文章讲述了一个人真的爱上了ChatGPT。

I had the CEO of Replica on the show a couple of weeks ago. She said that they are regularly invited to marriages of people who are marrying their replicas. They're moving into this more assistive space.

几周前，我邀请了Replica的首席执行官上节目。她说，他们经常被邀请参加那些与自己复制品结婚的人的婚礼。他们正迈向这种更具辅助性的领域。

Do you think that when we start interacting with something that knows us so well,  that helps us with everything we need, is it going to be like a third type of relationship where it's not necessarily a friend,

你认为，当我们开始与如此了解我们、帮助我们满足一切需求的存在互动时，这会不会演变成一种第三类型的关系，不一定是朋友，

 not a lover, but it's going to be a deep relationship, don't you think?

也不是情人，而是一种深层次的关系，你不这样认为吗？

Demis Hassabis:

Yeah, it's going to be really interesting. I think the way I'm modeling that, first of all, is at least two domains, first of all, which is your personal life and then your work life, right?

德米斯·哈萨比斯:

是的，这将非常有趣。我所构想的方式，首先至少涵盖两个领域，一是你的个人生活，二是你的工作生活，对吧？

So I think you'll have this notion of virtual workers or something. Maybe we'll have A set of them or managed by a lead assistant that does a lot of the,  helps us be way more productive at work, you know,

所以我认为你会有这种虚拟工作者的概念。也许我们会拥有一组这样的助手，或者由一位首席助手来管理，他们能大大提升我们的工作效率，

 or whether that's email across workspace or whatever that is. So we're really thinking about that.

无论是处理工作邮箱还是其他事务。所以我们正认真思考这一点。

Then there's the personal side where, you know, we're talking about earlier about all these booking holidays for you, arranging things, mundane things for you, sorting things out. And then that makes your life more efficient.

然后是个人方面，比如我们之前提到的为你预订假期、安排日程、处理琐碎事务、整理一切，这些都能让你的生活更高效。

I think it can also enrich your life, so recommend you things, amazing things that it knows you as well as you know yourself.

我认为它还能丰富你的生活，为你推荐那些它了解得和你了解自己一样透彻的精彩事物。

So those two I think are definitely going to happen and then I think there is a philosophical discussion to be had about is there a third space where these things start becoming so integral to your life they become more like companions.

所以我认为这两方面肯定会实现，然后我觉得还有一个哲学层面的讨论：是否存在第三种领域，在那里这些东西变得如此融入你的生活，以至于它们更像是伴侣。

I think that's possible too.

我认为这也是可能的。

We've seen that a little bit in gaming so you may have seen we had a little prototypes of Astro working in and Gemini working with like being almost a game companion commenting and you almost like as if you had a friend looking at a game you're playing and recommending things to you and advising you but also maybe just playing along with you and it's very fun.

在游戏领域我们已经稍微看到了一些这样的情况，你可能见过我们的一些原型，比如Astro和Gemini，它们几乎就像游戏中的伙伴，在评论，你仿佛有个朋友在观看你玩游戏、向你推荐和建议，同时也可能与你一同游戏，这非常有趣。

I haven't quite thought through all the implications of that, but they're going to be big. I'm sure there is going to be demand for companionship and other things.

我还没有完全理清其中所有的含义，但它们肯定会产生重大影响。我确信对伴侣般存在的需求以及其他需求都会大幅增加。

Maybe the good side of that is it will help with loneliness and these sorts of things,  but I think it's going to have to be really carefully thought through by society what directions we want to take that in.

也许这其中的好处在于它能缓解孤独感等问题，但我认为社会必须非常谨慎地考虑我们想朝哪个方向发展。

Speaker 1:

I mean my personal opinion is that that it's the most underappreciated part of AI right now and that people are just going to form such deep relationships with these bots as they get better because like I know it's a meme in AI that this is the worst it's ever going to be.

发言者 1:

我的个人观点是，目前这是人工智能中最被低估的部分，人们会随着这些机器人的不断进步而与它们建立如此深厚的关系，因为在AI界有个梗，说这已经是最糟糕的状态了。

Demis Hassabis:

Yeah.

德米斯·哈萨比斯:

是的。

Speaker 1:

And it's going to be crazy.

发言者 1:

这将会非常疯狂。

Demis Hassabis:

Yeah, I think I think it's going to be pretty crazy. This is what I meant about the under underappreciating what's to come. I still don't think this this kind of thing I'm talking about. Right. I think that it's going to be really crazy.

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

德米斯·哈萨比斯:

是的，我认为这会非常疯狂。这就是我所说的，未来会被严重低估。我依然认为我所谈论的这些事情会非常疯狂。

它将带来极大的颠覆。我认为其中也会有许多积极的方面，很多事情会变得惊人而美好，但同时我们也将面临进入这个全新勇敢世界的风险。

Speaker 1:

So you brought up Astra a couple times. Let's just talk about it. It's Project Astra, as you call it. It is almost an always-on AI assistant. You can hold your phone. It's currently just a prototype.

They're not publicly released, but you can hold your phone and it will see what's going on in the room. I can basically, I've seen you do this on your show or not you personally, but somebody on your team. You can say, okay, where am I?

And I'll be like, oh, you're in a podcast studio or anything. Okay. So it could have this contextual awareness.

发言者 1:

所以你提到了Astra好几次。让我们来聊聊它。这就是你所说的Project Astra，它几乎是一个始终在线的AI助手。你可以拿着你的手机。它目前只是一个原型。

它们还没有公开发布，但你可以拿着手机，它就能察觉到房间里发生的事情。我基本上见过你们团队中的某个人在节目中这样操作。你可以说，“好的，我现在在哪里？”

而它会回答，“哦，你在播客录音室或其他地方。”所以它具备情境感知能力。

Demis Hassabis:

Yes.

德米斯·哈萨比斯:

是的。

Speaker 1:

Can that work without smart glasses? Because it's really annoying to hold my phone up. So like when are we going to see Google smart glasses with this technology embedded?

发言者 1:

它能在没有智能眼镜的情况下工作吗？因为拿着手机实在太麻烦了。那么，我们什么时候能看到嵌入了这项技术的Google智能眼镜呢？

Demis Hassabis:

They're coming. So we teased it in some of our early prototypes. So that we're mostly prototyping on phones currently because they have more processing power. But where of course Google's always been a leader in glasses.

德米斯·哈萨比斯:

它们正在开发中。我们在一些早期原型中已经略微展示过这一点。目前我们主要在手机上进行原型测试，因为手机拥有更强的处理能力。当然，Google一直是智能眼镜领域的领导者。

Speaker 1:

Google Glass.

发言者 1:

谷歌眼镜。

Demis Hassabis:

Yeah, and exactly.

德米斯·哈萨比斯:

是的，正是如此。

Speaker 1:

Just a little too early.

发言者 1:

只是还太早了。

Demis Hassabis:

Yeah, maybe a little too early. And now I actually think, and they're super excited, that team is that You know, maybe this assistant is the killer use case that glasses has always been looking for.

And I think it's quite obvious when you start using Astra in your daily life, which we have with trusted testers at the moment and in kind of beta form, there are many use cases where it would be so useful to use it,

but it's inconvenient that you're holding the phone. So one example is while you're cooking, for example.

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

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

德米斯·哈萨比斯:

是的，也许确实还太早。现在我实际上认为，他们团队非常兴奋，认为也许这个助手正是智能眼镜一直在寻求的杀手级应用场景。

我觉得很明显，当你开始在日常生活中使用Astra时（目前我们在信任测试者中以测试版形式进行），你会发现有许多使用场景会非常有用，

但拿着手机使用实在不方便。举个例子，比如你在做饭时，

它可以建议你接下来该做什么，提供菜单，判断你是否正确切割或正确油炸了食材，而你希望它是免提的。

我认为在未来几年内，智能眼镜以及其他免提设备将大放异彩，我们计划站在这一领域的前沿。

Speaker 1:  

Other form factors?  

发言者 1:  

其他形态吗？

Demis Hassabis:  

Well, you could imagine earbuds with cameras and, you know, glasses is the obvious next stage. But is that the optimal form? Probably not either. But partly, we've also got to see,  

德米斯·哈萨比斯:  

嗯，你可以想象带有摄像头的耳塞，而眼镜显然是下一阶段。但那是最优的形态吗？可能也不是。不过部分原因是，我们还得看看，

we're still very early in this journey of seeing what are the regular user journeys and killer sort of use journeys that everyone uses,  bread and butter uses every day. And that's what the Trusted Tester program is for at the moment.  

我们在探索用户常规使用路径以及人人每天依赖的基本、杀手级应用的过程中还处于非常早期的阶段。而这正是当前Trusted Tester项目的目的。

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

我们正收集这些信息，观察人们的使用情况，并看看哪些功能最终会被证明是有用的。

Speaker 1:  

Okay, one last question on agents then we move to science. Agentic agents, AI agents, this has been the buzzword in AI for more than a year now.  

发言者 1:  

好，关于代理的问题最后一个，然后我们转到科学领域。代理性代理、AI代理，这已经是人工智能领域超过一年的流行词了。

Demis Hassabis:  

Yeah.  

德米斯·哈萨比斯:  

是的。

Speaker 1:  

There aren't really any AI agents out there.  

发言者 1:  

实际上还没有真正的AI代理。

Demis Hassabis:  

No.  

德米斯·哈萨比斯:  

没有。

Speaker 1:  

What's going on?  

发言者 1:  

这是怎么回事？

Demis Hassabis:  

Yeah. Well, again, you know, I think the hype train can potentially is ahead of where the actual science and research is. But I do believe that this year will be the year of agents, the beginnings of it.  

我想，再说一次，我认为炒作可能确实领先于实际的科学和研究。但我相信今年将是代理的元年，至少是一个开始。

I think you'll start seeing that, you know, maybe second half of this year. But there'll be the early versions. And then, you know, I think they'll rapidly improve and mature. So, but I think you're right.  

我认为你会在今年下半年看到这一现象，会有早期版本出现。然后，我觉得它们会迅速改进并成熟。所以，我同意你的看法。

I think the technology at the moment is still in the research lab, the agent technologies, but things like Astra, robotics, I think it's coming.  

我认为目前代理技术仍然停留在研究实验室阶段，但像Astra、机器人等技术，我觉得它们正在到来。

Speaker 1:  

Do you think people are going to trust them? I mean, it's like, go use the internet for me. Here's my credit card. I don't know.  

发言者 1:  

你认为人们会信任它们吗？我的意思是，它们就像是，“帮我上网，这是我的信用卡”，我不知道。

Demis Hassabis:  

Well, so I think to begin with, you would probably, my view at least, would be to not allow, have human in the loop for the final steps. Like, don't pay for anything. Use your credit card unless the human user operator authorizes it.  

德米斯·哈萨比斯:  

嗯，我认为至少在开始时，我的看法是，在最后环节中不要让人类介入。比如，不要让它们支付任何费用，除非得到人工操作员的授权，否则不要使用你的信用卡。

So that would, to me, be a sensible first step. Also, perhaps, certain types of activities or websites or whatever, kind of off limits, you know, banking websites and other things in the first phase.  

对我来说，这将是一个明智的第一步。此外，或许某些类型的活动或网站，比如银行网站等，也应该在第一阶段设为禁区。

While we continue to test out in the world that how robust these systems are.  

同时，我们会在实际应用中继续测试这些系统的稳健性。

  

Speaker 1:  

I propose we've really reached AGI when they say, don't worry, I won't spend your money. And then they do the deceptiveness thing. And then next thing you know, you're on a flight somewhere.  

发言者 1:  

我认为当它们说“别担心，我不会花掉你的钱”，然后又做出某种欺骗行为，接着你发现自己正乘坐着飞往某地的航班时，我们才真正达到了AGI。

Demis Hassabis:  

Yes. Yeah, that would be that would be that would be getting closer for sure. For sure. Yeah.  

德米斯·哈萨比斯:  

是的，那样肯定会更接近目标。绝对如此。

  

Speaker 1:  

All right, science. So you worked on basically decoding all protein folding with AlphaFold. You won the Nobel Prize for that.  

发言者 1:  

好了，谈谈科学。你用AlphaFold基本上破解了蛋白质折叠的问题，并因此赢得了诺贝尔奖。

Not to skip over the thing that you won the Nobel Prize for, but I want to talk about what's on the roadmap, which is that you have an interest in mapping a virtual cell.  

这并不是要跳过你因之获得诺贝尔奖的成就，而是我想谈谈你未来规划中的一个项目——虚拟细胞的构建。

Demis Hassabis:  

Yes.  

德米斯·哈萨比斯:  

是的。

Speaker 1:  

What is that and what does it get us?  

发言者 1:  

那是什么？它能为我们带来什么？

Demis Hassabis:  

Yeah. Well, so if you think about what we did with AlphaFold was essentially solve the problem of finding the structure of a protein. Proteins, everything in life depends on proteins, right? Everything in your body.  

德米斯·哈萨比斯:  

是的。那么，想想我们用AlphaFold做的事情，基本上就是解决了寻找蛋白质结构的问题。蛋白质——生活中的一切都依赖于蛋白质，对吧？你体内的一切亦然。

So that's the kind of static picture of a protein. The thing about biology is really it's you only understand what's going on in biology if you understand the dynamics and the interactions between the different things in a cell.  

这仅仅是蛋白质的静态图像。生物学的关键在于，你只有理解细胞内各个部分之间的动态和相互作用，才能真正明白生物学中发生了什么。

And so a virtual cell project is about building a simulation, an AI simulation of a full working cell. I probably start with something like a yeast cell because of the simplicity of the yeast organism. And you have to build up there.  

因此，虚拟细胞项目就是构建一个全功能细胞的模拟，一个由AI驱动的细胞模拟。我可能会从像酵母细胞这样的简单生物开始，然后逐步扩展。

So the next step is with AlphaFold3, for example, we started doing pairwise interactions between proteins and ligands and proteins and DNA proteins and RNA.  

接下来的步骤，比如在AlphaFold3中，我们开始研究蛋白质与配体、蛋白质与DNA、蛋白质与RNA之间的成对相互作用。

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

接着，下一步将是对整个生物通路建模，也许是某个癌症通路或类似的、对解决某种疾病有帮助的通路，最终建立一个完整细胞的模型。

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

这之所以重要，是因为你能够提出假设，并测试这些假设，比如改变某种营养成分或向细胞中注入药物，然后观察细胞如何响应。

And at the moment, of course, you have to do that painstakingly in a wet lab. But imagine if you could do it a thousand, a million times faster in silico first, and only at the last step do you do a validation in the wet lab.  

而目前，当然，这些都必须在湿实验室里费力地进行。但想象一下，如果你能在计算机模拟中快上千上百万倍地进行，然后只在最后一步在湿实验室中验证，那会怎样？

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

这样一来，你就可以将那部分在湿实验室中进行的、成本和时间上都高出数百万倍的搜索过程，转移到计算机模拟中进行。

So it's again, it's sort of translating, again, what we did in the games environments, but here in the sciences and the biology. So you, you build a model, and then you use that to do the reasoning and the search over,  

这再次类似于我们在游戏环境中所做的事情，不过这次是在科学和生物学领域。你构建一个模型，然后用它来进行推理和搜索，

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

然后得出的预测，至少比现有方法更好，也许并不完美，但足够有用，可以供实验人员进行验证。

Speaker 1:

And the wet lab is within people.

发言者 1:

而湿实验室的部分则由人来完成。

Demis Hassabis:

Yeah, so the wet lab, you'd still need a final step with the wet lab to prove what the predictions were actually valid. So, you know, but you wouldn't have to do all of the work to get to that prediction in the wet lab.

So you just get here's the prediction. If you put this chemical in, this should be the change, right? And then you just do that one experiment. So and then after that, of course, you still have to have clinical trials.

If you're talking about a drug, you would still need to test that properly through the clinical trials and so on and test it on humans for efficacy and so on. That, I also think, could be improved with AI, that whole clinical trial.

That also takes many, many years. But that would be a different technology from the virtual cell. The virtual cell would be helping the discovery phase for drug discovery.

德米斯·哈萨比斯:

是的，所以湿实验室这部分，你仍然需要在湿实验室进行最后一步，以证明预测实际上是有效的。也就是说，你不必在湿实验室里完成所有工作来得出那个预测。

你只需得到预测结果，比如“如果你加入这种化学物质，就会发生这种变化”，然后你只需做那一次实验。当然，之后如果涉及药物，你仍然需要通过临床试验来正确测试它的有效性等等。我也认为整个临床试验过程可以通过AI得到改进。

而且这也需要很多很多年。但这将是一种不同于虚拟细胞的技术。虚拟细胞将帮助药物发现阶段的研发。

Speaker 1:

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

发言者 1:

就像我有一个药物创意，把它放进虚拟细胞中试试。

Demis Hassabis:

See what it does. Yeah. And maybe eventually it's a liver cell or a brain cell or something like that. So you have different cell models. And then, you know, at least 90% of the time, it's giving you back what would really happen.

德米斯·哈萨比斯:

观察它的反应。是的，也许最终可以模拟成肝细胞、脑细胞或类似的细胞模型。所以你会有不同的细胞模型，而且至少90%的时间，它返回的结果会非常接近真实情况。

Speaker 1:

That'd be incredible. How long do you think that's going to take to figure out?

发言者 1:

那将是不可思议的。你觉得这需要多长时间才能实现？

Demis Hassabis:

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

德米斯·哈萨比斯:

我认为可能再过五年左右。

Speaker 1:

Okay.

发言者 1:

好。

Demis Hassabis:

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

德米斯·哈萨比斯:

是的。所以我有一个大约五年的项目，很多AlphaFold的前团队成员正在致力于这项工作。

Speaker 1:

Yeah. I was asking your team here. So you figured it out. Yeah. I speak with him.

发言者 1:

是的。我当时问了你们团队，所以你们搞定了。我和他谈过了。

Demis Hassabis:

Yeah.

德米斯·哈萨比斯:

嗯。

Speaker 1:

I was like, you figured out protein folding. What's next? And this is like, it's just very cool to hear about these new challenges because yeah, developing drugs is a mess right now. We have so many promising ideas.

They never get out the door because Just the process is absurd.

发言者 1:

我当时就想，你们已经破解了蛋白质折叠问题，接下来是什么？听到这些新挑战真是太酷了，因为现在开发药物真是一团糟。我们有那么多有前景的创意，但它们从来无法顺利推进，因为整个流程太荒谬了。

Demis Hassabis:

It's process too slow and discovery phase too slow. I mean, look how long we've been working on Alzheimer's and it's a tragic way for someone to go and for the families and we should be a lot further. It's 40 years of work on that.

德米斯·哈萨比斯:

整个流程太慢，发现阶段也太慢。想想我们在阿尔茨海默症上已经研究了多久，这对患者和家庭来说真是悲剧，我们本应取得更大进展。这已经花了40年的时间。

Speaker 1:

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

发言者 1:

是的，我家里就见过几次这种情况，如果我们能确保这种情况不再发生，那就太好了……

Demis Hassabis:

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

德米斯·哈萨比斯:

在我看来，这可能是我们可以用AI实现的最有价值的事情之一。

Speaker 1:

Yeah, it's a terrible way to see somebody decline. So, it's important work. On addition to that, there's the genome and so the Human Genome Project sort of, I was like, okay, so they decoded the whole genome.

There's no more work to do there, like just same way that you decoded proteins with fold, but it turns out that Actually, we just have like a bunch of letters when it's decoded.

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

发言者 1:

是的，看着某人衰退真是太可怕了。所以这项工作非常重要。除此之外，还有基因组项目，我当时想，“好吧，他们解码了整个人类基因组。”

结果发现，其实那只是一些字母而已，就像你用AlphaFold解码蛋白质那样，但实际上，我们得到的只是一串字母。

所以现在你正致力于用AI来解读这些字母的含义？

Demis Hassabis:

Yes. So yeah, we have lots of cool work on genomics and trying to figure out if mutations are going to be harmful or benign, right? Most mutations to your DNA are harmless. But of course, some are pathogenic.

And you want to know which ones there are. So our first systems are the best in the world at predicting that.

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

And a lot of more complex diseases that we haven't made progress with are probably not due to a single mutation. That's more like rare childhood diseases, things like that.

So there, you know, we need to, I think AI is the perfect tool to sort of try and figure out what these weak interactions are like, right? How they maybe kind of compound on top of each other.

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

德米斯·哈萨比斯:

是的。所以，我们在基因组学领域有很多很酷的工作，试图判断基因突变是有害的还是无害的，对吧？大多数DNA突变都是无害的，但当然，有些是致病的。

而你需要知道哪些突变存在。所以我们的第一个系统在预测这一点方面是世界上最优秀的。

而且许多我们尚未取得进展的更复杂疾病，可能并非由于单一突变引起，而更像是罕见的儿童疾病之类的。

因此，我们需要……我认为AI是试图弄清这些微弱相互作用的完美工具，对吧？看它们如何可能相互叠加。

也许这些统计数据并不十分明显，但一个能够识别模式的AI系统将能够发现这里存在某种联系。

Speaker 1:

And so we talk about this a lot in terms of disease, but also I wonder what happens in terms of making people superhuman. I mean, if you're really able to tinker with the genetic code, right, the possibilities seem endless.

发言者 1:

所以我们经常讨论疾病问题，但我也在想，关于让人类变得超级人类会发生什么。我是说，如果你真的能改造基因密码，其可能性似乎无穷无尽。

Speaker 1:

So what do you think about that? Is that something that we're going to be able to do through AI?

发言者 1:

那你怎么看？这会是我们能够通过人工智能实现的事情吗？

Demis Hassabis:

I think one day. I mean, we're focusing much more on the disease profile.

德米斯·哈萨比斯:

我认为总有一天会实现。我的意思是，我们目前更多地关注于疾病的方面。

Speaker 1:

That would be the first step.

发言者 1:

那将是第一步。

Demis Hassabis:

Yeah, that's the first step. And I've always felt that that's the most important. If you ask me what's the number one thing I wanted to use AI for and the most important thing we use AI for is for helping human health.

But then, of course, beyond that, one could imagine aging, things like that. Of course, there's a whole field in itself. Is aging a disease? Is it a combination of diseases? Can we extend our healthy lifespan?

These are all important questions and I think very interesting and I'm pretty sure AI will be extremely useful in helping us find answers to those questions too.

德米斯·哈萨比斯:

是的，那是第一步。我一直觉得这最为重要。如果你问我最想用人工智能做什么，我会说最重要的用途就是帮助改善人类健康。

但是，当然，除此之外，你也可以设想延缓衰老之类的事情。当然，这本身就是一个完整的领域。衰老是疾病吗？还是多种疾病的组合？我们能延长健康寿命吗？

这些都是非常重要且有趣的问题，我确信人工智能在帮助我们找到这些问题的答案方面也会极为有用。

Speaker 1:

I see memes come across my Twitter feed and maybe I need to change the stuff I'm recommended, but it's often like if you will live to 2050, you're not going to die. What do you think the potential max lifespan is for a person?

发言者 1:

我在推特上常看到一些梗，也许我得换换推荐的内容，但常见的说法是，如果你能活到2050年，你就不会死。你认为一个人的潜在最长寿命是多少？

Demis Hassabis:

Well, look, I know a lot of those folks in aging research very well. I think it's very interesting the pioneering work they do. I think there's nothing good about getting old and your body decaying.

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

And so I think anything we can alleviate human suffering and extend healthy lifespan is a good thing. You know, the natural limit seems to be about 120 years old.

But from what we know, you know, if you look at the oldest people that are lucky enough to live to that age. So there's, you know, it's an area I follow quite closely.

I don't have any, I guess, new insights that are not already known in that. But I would be surprised if that's the limit, right? There's a sort of two steps to this.

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

But then that's not enough to probably get you past 120, because there's some sort of then there's the question of just natural systemic decay. Right, aging, in other words, so not specific disease, right?

Often those people that live to 120, they don't seem to die from a specific disease. It's just sort of just general atrophy. So then you're going to need something more like rejuvenation, where you rejuvenate your cells or you,

you know, maybe stem cell research, you know, companies like Altos are working on these things, resetting the cell clocks. It seems like that could be possible.

But again, I feel like it's so complex because biology is such a complicated emergent system. In my view, you need AI to be able to crack anything close to that.

德米斯·哈萨比斯:

嗯，你看，我非常了解那些从事衰老研究的人。我觉得他们做的开创性工作非常有趣。我认为变老和身体衰退没有任何好处。

我想，如果有人亲眼见过亲属经历这一过程，无论对家庭还是对个人来说，这都是一件非常艰难的事，对吧？

所以我认为，任何能减轻人类痛苦、延长健康寿命的事情都是好事。你知道，自然寿命的极限似乎大约在120岁左右。

但据我们所知，如果你看看那些幸运地活到那个年龄的长寿者……这也是我一直密切关注的领域。

我没有什么全新的见解，但如果那真的是极限，我会感到惊讶。这个问题大致可以分为两个步骤。

第一步是有朝一日治愈所有疾病，我认为我们将通过Isomorphic以及我们正在进行的分拆项目、我们的药物发现分拆项目来实现这一点。

但那可能还不足以让你超过120岁，因为接下来还有一个问题，就是自然的系统性衰退。换句话说，衰老不仅仅是某种特定疾病的问题，对吧？

那些能活到120岁的人通常并不是死于某种特定疾病，而是因为整体机能衰退。所以你可能需要一些更像是细胞再生的手段，比如让你的细胞恢复活力，或者，你知道，也许像干细胞研究这样的领域，诸如Altos这样的公司正在做这些事情，重置细胞时钟。看起来这是可能的。

但我觉得这非常复杂，因为生物学是一个如此复杂的涌现系统。在我看来，要解决这类问题，你需要依靠人工智能。

Speaker 1:

Very quickly on material science, I don't want to leave here without talking about the fact that you've discovered many new materials or potential materials. The stat I have here is known to humanity.

Recently, there were 30,000 stable materials. You've discovered 2.2 million with a new AI program. Just dream a little bit, because we don't know what all those materials can do.

We don't know whether they'll be able to handle being out of like a frozen box or whatever. Dream materials for you to find in that set of new materials.

发言者 1:

接下来快速聊聊材料科学，我不想不提你发现了许多新材料或潜在材料这一事实。据说目前已知的稳定材料只有3万个，而你们用一个新的人工智能程序发现了220万个。稍微放飞一下想象力，因为我们不知道这些材料究竟能做些什么。

我们也不清楚它们是否能适应从冰箱中取出的环境，或者其他什么条件。那些梦幻般的材料就藏在这220万个新材料中，等待你去发现。

Demis Hassabis:

Well, I mean, we're working really hard on materials. To me, it's like one of the next sort of big impacts we can have, like the level of alpha fold really in biology, but this time in chemistry and materials.

You know, I dream of one day discovering a room temperature superconductor.

德米斯·哈萨比斯:

嗯，我是说，我们在材料科学方面也在非常努力地工作。对我来说，这就像在生物学中AlphaFold带来的巨大影响力，但这次是在化学和材料领域。

你知道，我梦想有一天能发现室温超导体。

Speaker 1:  

So what will that do? Because that's another big meme that people talk about.  

发言者 1:  

那会有什么作用呢？因为那是人们经常谈论的另一个热门话题。

Demis Hassabis:  

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

德米斯·哈萨比斯:  

嗯，它将有助于解决能源危机和气候危机，因为如果你拥有廉价的超导体，就可以把能量从一个地方传输到另一个地方而不会损失。

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

没错，这样你就可以在撒哈拉沙漠安装太阳能板，然后利用超导体将能量输送到欧洲等需要的地方。

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

目前，传输过程中会损失大量能量，转化为热能和其他形式的能量损耗。

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

因此，你还需要其他技术，比如电池之类的储能装置，因为你无法将能量以极高的效率直接传送到目标地点。

So, but also materials could help with things like batteries too, like come up with the optimal battery. I don't think we have the optimal battery designs. That maybe we can do things like a combination of materials and proteins.  

另外，新材料也可以帮助改进电池，比如研发出最优电池。我认为我们目前还没有最优的电池设计，也许我们可以通过结合材料和蛋白质来实现。

We can do things like carbon capture, you know, modify algae or other things to do carbon capture better than our artificial systems. I mean, even the one of the most famous and most important chemical processes,  

我们还可以开展类似碳捕捉的工作，比如改造藻类或其他生物，使其在碳捕捉方面优于现有的人工系统。我的意思是，甚至连最著名、最重要的化学过程之一，

the Haber process to make fertilizer and ammonia, you know, to take nitrogen out of the air,  was something that allows modern civilization.  

哈柏法（用来制造肥料和氨，将空气中的氮固定下来）的发明，正是支撑现代文明的重要技术。

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

但如果我们知道正确的催化剂和合适的材料，可能还有许多其他化学过程也能通过这种方式得到催化。

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

因此，我认为最具影响力的技术之一将是利用计算机模拟（in silico）进行材料设计。

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

我们已经完成了第一步，证明了我们可以设计出新的稳定材料，但我们还需要一种方法来测试这些材料的性能。

There's no lab can test 200,000, you know, tens of thousands of materials or millions of materials at the moment. So we have to, that's the hard part is to do the testing.  

Speaker 1:  

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

发言者 1:  

你认为那个室温超导体就在这其中吗？

Demis Hassabis:  

Well, I heard that we actually think there are some superconductor materials. I doubt they're room temperature ones though, but I think at some point, if it's possible with physics, an AI system will one day find it.  

德米斯·哈萨比斯:  

嗯，我听说我们实际上认为存在一些超导材料，虽然我怀疑它们是室温超导体，但我认为如果物理上可能，总有一天AI系统会发现它的。

Speaker 1:  

So that's one use. The two other uses I could imagine, probably people interested in this type of work, toy manufacturers and militaries. Are they working with it?  

发言者 1:  

所以这是一种应用。我还能想象另外两种用途，可能对这类技术感兴趣的人包括玩具制造商和军方。他们是否也在利用这项技术呢？

Demis Hassabis:  

Yeah, I mean, look, I think there is incredible one. I mean, a big part of my early career was in game design. Yeah, theme park and simulations. That's what got me into simulations and AI in the first place.  

德米斯·哈萨比斯:  

是的，我的意思是，看，我认为这领域潜力巨大。我的早期职业生涯中很大一部分时间是在游戏设计、主题公园和模拟领域工作，这正是最初让我进入模拟和人工智能领域的原因。

And I've always loved both of those things. And in many respects, the work I do today is just an extension of that. And I just dream about like, what could I have done?  

而且我一直热爱这两者。在很多方面，我今天所做的工作只是对那段经历的延伸。我常常幻想，如果25-30年前我能拥有今天的AI技术，我能做些什么？

What kinds of amazing game experiences could have been made if I'd had the AI I have today, available 25-30 years ago when I was writing those games? And I'm a little bit surprised the game industry hasn't done that.  

如果当时我拥有今天这样的AI技术，在写那些游戏时能打造出怎样令人惊叹的游戏体验？我有点惊讶，游戏产业居然还没有这么做。

Speaker 1:  

We're starting to see some crazy stuff with NPCs that, like, are starting to...  

发言者 1:  

我们开始看到一些疯狂的现象，比如NPC们似乎开始……  

Demis Hassabis:  

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

德米斯·哈萨比斯:  

是的，NPC们当然存在，但更重要的是会出现智能的、动态的故事线。而且还会有全新类型的首创AI游戏，其角色和代理能够学习。

I once worked on a game called Black and White where you had a creature that you were nurturing. It was a bit like a pet dog that learned what you wanted. But we were using very basic reinforcement learning. This was like in the late 90s.  

我曾经参与开发一款名为《黑与白》的游戏，里面你要培养一个生物，有点像一只宠物狗，会学会你所期望的。但当时我们用的是非常基础的强化学习，这大概是在90年代末期。

Imagine what could be done today. And I think the same for maybe smart toys as well. And then of course on the militaries, you know, unfortunately, AI is a dual purpose technology.  

想象一下，如今能做到什么。而且我认为智能玩具也是如此。当然，对于军方而言，不幸的是，人工智能是一项双刃剑技术。

So one has to confront the reality that, especially in today's geopolitical world, people are using some of these general purpose technologies to apply to drones and other things. And it's not surprising that that works.  

因此，我们不得不面对这样一个现实，尤其在当今的地缘政治环境下，人们正利用这些通用技术应用于无人机等领域，而这种做法的有效性并不令人惊讶。

Speaker 1:  

Are you impressed with what China's up to? I mean, DeepSeek is this new model getting?  

发言者 1:  

你对中国在这方面的进展印象深刻吗？我是说，DeepSeek这个新模型怎么样？

Demis Hassabis:  

It's a little bit unclear how much they relied on Western systems to do that, you know, both training data, there's some rumors about that, and also maybe using some of the open source models as a starting point.  

德米斯·哈萨比斯:  

目前还不太清楚他们在多大程度上依赖了西方系统，比如训练数据上有一些传闻，还有可能是以一些开源模型作为起点。

But look, it's for sure, it's impressive what they've been able to do. And, you know, I think that's something we're going to have to think about how to keep The Western frontier models in the lead, I think they still are at the moment,  

但你看，毫无疑问，他们所能做到的确实令人印象深刻。而且，我认为我们必须思考如何让西方前沿模型继续保持领先地位，我认为目前它们依然领先，

but you know for sure China's very very capable at engineering and scaling.  

但可以肯定的是，中国在工程和规模化方面非常非常强大。

Speaker 1:  

Let me ask you one final question. Just give us your vision of what a world looks like when there's superintelligence. Let's move past, we started with AGI, let's end on superintelligence.  

发言者 1:  

让我问你最后一个问题。请描述一下当超智能出现时，世界会是什么样子。我们从AGI开始，现在谈谈超智能吧。

Demis Hassabis:  

Yeah, well look, I think for there, two things there. One is I think a lot of the best sci-fi we can look at as interesting models to debate about what kind of galaxy or universe do we want to, a world do we want to move towards.  

德米斯·哈萨比斯:  

嗯，我认为在这方面有两点。首先，我觉得我们可以参考许多优秀的科幻作品，作为讨论我们希望迈向何种银河或宇宙、何种世界的有趣模型。

And the one I've always liked most is actually the Culture series by Ian Banks. I started reading that back in the 90s and I think that is a picture, it's like a thousand years into the future,  

而我一直最喜欢的就是伊恩·班克斯的《文化》系列。我在90年代就开始阅读它，我认为它描绘了未来大约一千年后的景象，

but it's in a post-AGI world where there are AGI systems coexisting with human society and also alien society and we've Humanity has basically maximally flourished and spread to the galaxy.  

那是一个后AGI时代，AGI系统与人类社会甚至外星社会共存，而人类基本上达到了极致的繁荣，并扩散到了整个银河系。

And that I think is a great vision of how things might go in the positive case. So I sort of hold that up.  

我认为这是对未来积极情形的一种极佳设想。所以我一直抱有这种愿景。

I think the other thing we're going to need to do is as I mentioned earlier about the under-appreciating still what's going to come in the longer term, I think there is a need for some great philosophers to, you know, where are they?  

我认为我们还需要做的另一件事是，正如我之前提到的，对于长期将要到来的事物仍被低估这一现状，我们需要一些伟大的哲学家来——你知道，他们都在哪儿呢？

The great next philosophers, the equivalents of Kant or Wittgenstein or even Aristotle.  

下一代伟大的哲学家，相当于康德、维特根斯坦甚至亚里士多德。

I think we're going to need that to help navigate society to that next step because I think the, you know, AGI and artificial superintelligence is going to change humanity and the human condition.  

我认为我们需要这些思想家来引领社会迈向下一个阶段，因为我相信AGI和人工超智能将会改变人类和人类的生存状态。

Speaker 1:  

Demis, thank you so much for doing this. Great to see you in person. I hope to do it again soon.  

发言者 1:  

德米斯，非常感谢你接受采访。见到你真是太棒了，希望不久的将来能再有机会聊聊。

Demis Hassabis:  

Thank you. Thank you very much.  

德米斯·哈萨比斯:  

谢谢你，非常感谢。

Speaker 1:  

Everybody, thank you for listening and we'll see you next time on Big Technology Podcast.  

发言者 1:  

各位听众，非常感谢收听，我们下次在《大科技播客》再见。