AI 解决过的最难的问题

关于智能的定义显然有些不对劲。如果我们推演下去，极限在哪里？AI 的最佳用途是改善人类健康。那是我一直等待的时刻，能实现其他系统无法做到的事情。我想用 AI 作为工具来帮助我们理解周围的现实本质。政府将会使用 AI。你希望他们用它来做什么？有两件事需要担心。那就是 Demis Hassabis，Google DeepMind 的 CEO。诺贝尔奖得主。他是当今在世最重要的人物之一，面对的是正在迅速成为我们一生中最大的技术飞跃。因为 AI 影响我们生活的最主要方式不是我们能看到的。它不是聊天机器人，也不是图像生成器。它是那些我们看不见的工具，用于药物设计、自然灾害检测、核聚变和量子计算。这些工具是他和他的团队正在构建的。这是他因其中一种工具获得诺贝尔奖的画面。所以，他是谁以及他选择构建什么，对你我都至关重要。他非常迷人。他小时候是国际象棋神童，17 岁时拒绝了据称来自游戏公司的百万美元工作机会，去上大学，然后获得了认知神经科学博士学位。他创立了 DeepMind 公司，使命是解决智能问题，从击败电子游戏开始。然后他把公司卖给了 Google，专门因为他们承诺让 DeepMind 专注于科学研究。但是，随着这演变成近代史上最激烈的技术之战，Demis 现在负责的事情要多得多。他现在基本上掌管 Google 在 AI 领域的一切。他做出的决定每天影响着你和数百万人的生活。那么，他打算用所有这些权力做什么？我的目标是向你展示 Demis Hassabis 想要构建的未来，这样你就可以自己判断它的好坏。欢迎来到《深度对话》。

Something's obviously not quite right about the definition of intelligence. If we play this out, what's the limit here? The best use case of AI was to improve human health. It was the moment I've been waiting for that could achieve something no other system could. I want to use AI as a tool to help us understand the nature of reality around it. Governments are going to use AI. What would you hope that they use it for? There's two things to worry about. One is that's Demis Hassabis, the CEO of Google DeepMind. Nobel Prize winner. He is one of the most important people alive on what is quickly becoming the biggest technological leap in our lifetime. Because the biggest way that AI is going to impact our lives isn't something that we can see. It's not a chatbot, it's not an image generator. It's tools that are invisible to us in drug design and natural disaster detection and nuclear fusion and quantum computing. Tools that he and his team are building. Here he is winning the Nobel Prize for just one of those tools. So, who he is and what he chooses to build matters a lot for you and me. And he's fascinating. He's a childhood chess prodigy who at 17 turned down a reportedly million-dollar job offer from a gaming company to go to college instead and then got a PhD in cognitive neuroscience. He founded his company DeepMind with a mission to solve intelligence starting with beating video games. He then sold that company to Google specifically because they promised to let DeepMind focus on scientific research. But, as this has turned into the most intense technological battle in recent history, Demis is now in charge of much, much more. He's now behind basically everything Google does in AI. He's making decisions that affect your life and millions of other lives every single day. So, what is he planning to do with all of that power? My goal is to show you the future that Demis Hassabis wants to build so that you can decide for yourself what you think of it. Welcome to Huge Conversations.

非常感谢你参加这个节目。太棒了。

Thanks so much for doing this. It's great.

很荣幸。

Appreciate it.

你已经知道《深度对话》是一种不同的采访。我不会问你财务问题，也不会问你管理风格。这些在其他地方都有很好的报道。我希望在这次对话中，我们更像是在一起制作一个解说视频。我准备了一些道具。这本来不是要玩叠叠乐。每个积木代表一个项目或一个模型，我想谈谈它们以及它们如何组合在一起。所以它们本来是视觉辅助工具，但我们在布置的时候开始用它们玩叠叠乐，结果比我想象的任何事情都有趣得多。而且，我知道你喜欢游戏。

You already know that Huge Conversations is a different kind of interview. I'm not going to ask you about financials, I'm not going to ask you about your management style. All well covered elsewhere. What I'm hoping to do in this conversation is think about it more like an explainer that we're making live together. And I have some props. This was not actually meant to be a Jenga game. Each block represents a project or a model and I want to talk about them and how they fit together. And so they were meant to be visual aids, but as we were setting up, we started playing Jenga with them and it turned out to be way more fun than anything I had planned. Also, I know that you like games.

是的，我喜欢游戏。所以这很棒。至少在采访中是第一次。

Yes, I love games. So, this is great. First in an interview anyway.

所以，我希望在这次对话中一起制作这个解说，帮助人们看到 AI 现在真正在发生什么，以及你看到的未来是什么？你希望在这次对话中做什么？

So, my hope in this conversation is to make this explainer together and to help people see what's happening right now in AI really and what is the future that you see coming? What are you hoping to do in this conversation?

是的，我 30 多年前进入 AI 领域的很多原因是为了推动科学和医学的发展，我一直认为 AI 可能是做到这一点的终极工具。所以，我希望我们今天能谈谈这个，这确实是我对 AI 应用的热情所在。当然它也可以应用于许多其他事情。哦，这会很有趣。

Yeah, a lot of the reasons that I got into AI 30-plus years ago now is to advance science and medicine and I've always thought of AI as potentially the ultimate tool to do that. So, I'm hoping we're going to talk about that today and really that's been my passion for what to apply AI to. But of course it can be applied to many things. Oh, this is going to be a lot of fun.

所以，在我们这个叠叠乐游戏中，很多积木是人们听说过的，对吧？这些是，你知道，这个是 Gemini。

So, in this Jenga game that we have, a lot of these are blocks that people will have heard of, right? These are, you know, this one is Gemini.

对。

Right.

但我认为，AI 有意义地塑造人们生活的方式，大多数时候是那些他们看不见的东西。所以，我想从你获得诺贝尔奖的项目开始谈。AlphaFold。叠叠乐玩得好。我想讲述 AlphaFold 的故事，包括所有的戏剧性，因为有些人可能没听过。但之后我想很快进入这类科学的前沿。为什么你决定从这么多问题中解决这个？

But I would argue that the ways in which AI is meaningfully shaping people's lives most are the things that are invisible to them most of the time. So, I want to start by talking about the project that you won the Nobel Prize for. AlphaFold. Good Jenga play. I want to tell the story of AlphaFold with all of its drama because some people might not have heard it. But then I want to get really quickly to the cutting edge of this sort of category of science. Why did you decide to tackle this problem out of all of the many?

嗯，我实际上是在剑桥读本科时偶然遇到的。我有很多生物学家朋友，其中一个特别痴迷于所谓的蛋白质折叠问题。蛋白质是你身体里一切所依赖的。它们使生物学成为可能，使生命成为可能。它们重要的是三维结构。所以，在体内它们折叠成三维结构，这些结构决定了它们的功能或部分功能。所以，蛋白质折叠问题实际上是关于能否仅从一维氨基酸序列预测这个三维结构。那是蛋白质折叠 50 年的大挑战。我喜欢挑战，喜欢谜题。所以从科学角度看，我无法抗拒，因为这可能被描述为生物学中的费马大定理。谁能不感兴趣呢？而且，当我第一次听说时，我认为这类问题有一天会适合 AI。尽管这是在 90 年代末，我们没有任何 AI 可以处理这个问题。但我认为有一天会可能的。最后，如果你解决了它，它会带来巨大的影响，因为它会为研究打开所有下游可能性，特别是在药物发现和理解疾病等方面。所以，我认为 AI 最重要的应用是改善人类健康。这对人类健康如此重要的原因是，直到现在，为了开发新药，我们必须花费数十万美元和数年的人力，通过 X 射线照射来找出单个蛋白质的结构。所以，我们已经找出了一些蛋白质结构，但速度慢且昂贵。

Well, I came across it actually as an undergrad in Cambridge. So, I had a lot of biologist friends and one of them specifically was obsessed with what's called the protein folding problem. So, proteins are what everything in your body relies on. They make biology possible and living possible. And what's important about them is their 3D structure. So, in the body they fold up into kind of 3D structures and those structures determine what function they have or partially determine what function they have. And so, the protein folding problem is really about can you predict this 3D structure just from the one-dimensional amino acid sequence? So, that's the kind of 50-year grand challenge of protein folding. So, I love challenges, I love puzzles. So, I couldn't resist it from a scientific point of view as this probably, you know, is described to me as the equivalent of Fermat's Last Theorem but for biology. So, who couldn't be interested in that? But also, when I first heard it, I thought the kind of problem it was would be suitable for AI one day. Even though we of course this is in the late '90s, we didn't have any kind of AI that would be possible to work on this. But I thought one day that would be possible. And then the final thing was just the impact it would make if you cracked it because it would open up all these downstream possibilities for research and especially in things like drug discovery and understanding disease. So, which I think is, you know, the most important thing to apply AI to is improving human health. And the reason that this would be huge for human health is that up until now, in order to develop new medicines, we'd have to spend hundreds of thousands of dollars and years of human effort to find out the structure of a single protein by shooting X-rays at it. So, we had figured out some protein structures, but it was slow and expensive.

所以，我跳过了你和你的团队所做的巨大努力。但我认为通过我问问题的方式，人们很明显知道你解决了它。所以，有一个时刻你意识到它真正有用，你解决了被称为现代医学中最重要的未解决问题之一。那是 2021 年。你在一个会议上。我很高兴那个会议有摄像头，因为那是我见过的最不可思议的时刻之一。我们可以用 AlphaFold 来解决它吗？我想你在和你的团队讨论建立一个系统，科学家可以发送请求来获取特定蛋白质，比如一个网站，然后得到折叠的蛋白质。

So, I'm skipping over an enormous amount of hard work here by you and your team. But I think by the way that I'm asking the questions, it is very obvious to people that you solved it. So, there's this moment where you realize that it is genuinely useful and you have solved what had been called one of the most important unsolved problems in modern medicine. And it's 2021. You're in a meeting. I am so glad that there was a camera in this meeting because it is one of the most incredible moments I have ever seen. Can we use AlphaFold to solve it? I think you're talking with your team about setting up a system where scientists could send in a request for a specific protein, like a website, and then get the protein folded.

然后另一个人有一个非常不同的想法。是的。你能带我回顾一下那次会议上发生了什么吗？然后你的反应令人难以置信，我真的很想知道你在想什么。

And then someone else has a very different idea. Yes. Can you walk me through what happens in that meeting? And then your reaction is incredible and I really want to know what you were thinking.

是的，当然。嗯，你看，有趣的是摄像机刚好在那次会议上。那天真是疯狂，他们很少跟着我们，但那次会议他们来了。通常，对于这类预测模型，传统做法是搭建一个服务器，然后其他科学家把他们的蛋白质序列发给你，说「我对这个蛋白质感兴趣，你能把预测结构发给我吗？」整个领域过去 40 多年都是这样做的。原因是大多数预测算法相当慢。可能需要几天时间，然后你通过电子邮件把结构发回去，再请求下一个。但我在那次会议上意识到，我们折叠蛋白质的速度有多快——不仅准确，而且快，只需几秒钟。我就在心里粗略估算：已知科学、已知自然中的蛋白质有多少？2 亿个。我们有多少台计算机？需要多少台？如果每 10 秒折叠一个，我在那次会议中间摆弄手机时意识到，一年内就能完成。那么，何必费劲搭建服务器、数据库和电子邮件客户端呢？我们可以自己折叠所有蛋白质——任何人可能请求和想要的任何蛋白质——然后把它放在某个数据库里，免费供全世界科学家使用。我突然想到：我们就该这么做。

Yeah. Sure. Well, look, it was funny that the cameras were there happened to be in that particular meeting. It's crazy that was that day, they very rarely followed us, but it was for that meeting. Normally, for these prediction models, the traditional thing is you set up a server and then other scientists send you their protein sequences and say, 'I'm interested in this protein. Can you send me back the predicted structure?' That's how it's been done in the whole field for the last 40-plus years. The reason is that most prediction algorithms are quite slow. Maybe it would take a few days and then you'd get back the structure by email, and then you'd ask for the next one. But once I realized in that meeting how quickly we could fold the proteins—not only how accurately but how quickly, in a matter of seconds—I was doing a back-of-the-envelope calculation: how many proteins are known to science, known in nature? 200 million. How many computers do we have? How many would we need? If we folded one every 10 seconds, I realized in the middle of that meeting while fiddling on my phone that it would be possible in a year. So, why go to all the effort of building servers, databases, and email clients when we could just fold everything ourselves—everything anyone could ever request and ever want—and then put it on a database somewhere for free for all scientists in the world to use? It just suddenly hit me: we should just do that.

现在你真的可以——我们为什么不直接这么做呢？嗯，所以这是一个选项。我们应该运行所有存在的蛋白质。然后发布它。突然，所有这些事情一定在我脑海中进行，我突然意识到那将是显而易见的事情，而且实际上可能比搭建服务器更省力。所以，实际上它会节省我们的时间。而在那次会议上，你的反应是「我们为什么不直接这么做？那会好得多。我们显然应该这么做。」然后你就做了。突然之间，这个曾经如此艰难的关键过程变得又快又容易。全世界的科学家都在使用它。这个巨大的未解问题现在解决了。说我们现在已经预测了几乎所有已知科学蛋白质的结构，对吗？

Now you can really— Why don't we just do that? Well, so that's one of the options. We should just run every protein in existence. And then release that. Suddenly, all these things must have been going on in the back of my mind and I suddenly realized that that would be the obvious thing to do and it would be actually probably less effort than standing up the server. So, actually it would save us time. And you in that meeting, your reaction is something like, 'Why don't we just do that? That would be way better. We should clearly do that.' And then you do. All of a sudden this crucial process that had been so hard is suddenly fast and easy. And it's being used by scientists all over the world. This huge unsolved problem now solved. Is it correct to say that we have now predicted the structure of almost all proteins known to science?

是的。而且我们不断更新它。所以，每当有人从海洋某处舀起一桶水，那桶海水里有很多不同类型的生物，然后他们测序所有生物。自人类基因组测序以来，测序技术已经提升了多个数量级。所以，问题是结构生物学——找到这些 3D 结构——远远落后于基因测序。现在有了像 AlphaFold 2 这样的计算资源，我们可以跟上：「哦，这里有一百万个来自我们发现的新奇生物的新基因序列；哦，这里是结构。」我们在欧洲生物信息学研究所有一个小团队，每年都会更新当年发现的所有新序列。所以，我们现在始终处于前沿。我们知道所有这些不同蛋白质结构大致是什么样子。

Yes. And we keep updating it. So, every time somebody scoops a pail out of the ocean somewhere and there are loads of different types of organisms in that bucket of seawater, and then they sequence them all. The sequencing technology has improved many orders of magnitude since the human genome was sequenced. So, the problem was structural biology—finding these 3D structures—was far behind genetic sequencing. Now with computational resources like AlphaFold 2, we can keep up with, 'Oh, here's a new million genetic sequences from some new strange organisms we found; oh, here are the structures.' We have a small team at the European Bioinformatics Institute that keeps updating every year all the new sequences found that year. So, we're now always at the cutting edge. We know what all these different protein structures mostly look like.

太棒了。

That's so awesome.

这确实非常了不起。尤其对于研究较冷门生物或动物的研究人员来说，比如小麦。我发现很多植物的基因组数据比哺乳动物和人类多得多，这很奇怪。它们似乎有多份基因组拷贝等等。植物世界是一个奇怪而奇异的世界。但我的植物科学家朋友们，他们没有像人类基因组那样的资源——人类基因组已经做了很多工作。但一些对人类社会仍然非常重要的较冷门生物，比如农作物，现在我们能够直接跳到他们想用蛋白质做什么的科学问题上，也许帮助它们更好地适应气候变化。他们可以直接跳到他们真正感兴趣的问题上，而不是陷入试图结晶他们感兴趣的蛋白质的困境中。

It is pretty amazing. It's especially amazing for researchers that work on slightly more obscure organisms or animals, like wheat. I found out that a lot of plants have way more genomic data than mammals and humans, which is very strange. They seem to have multiple copies of their genome and things. It's a kind of strange and bizarre world, the plant world. But my plant scientist friends, they don't have the resources like with the human genome—a lot of work has been done on that. But some of these more obscure organisms that are still really important for humanity, like crops, now we're able to immediately jump to the science around what they want to do with the proteins, maybe help them be more resilient to climate change. They can jump straight to the problem they're actually interested in rather than getting bogged down with trying to crystallize the proteins they're interested in.

另一个好处是针对研究被忽视疾病的研究人员，这些疾病主要影响发展中国家，比如疟疾、恰加斯病或利什曼病。这些疾病影响着全球数亿人。但大型制药公司如果研究这些并寻找治疗方法，并没有太多利润，因为这些疾病发生在较贫困地区。所以，相关研究往往被忽视。有一些很棒的非营利组织在做这些研究。但他们没有很多资金或资源，所以给他们提供涉及疟疾病毒等蛋白质的结构，对他们来说也是巨大的福音，因为他们可以直接进入药物发现阶段。这是我做这项研究时最难弄清楚的事情之一，因为有一个时刻，全世界的科学家都可以使用 AlphaFold。你可以看到地图亮起来。你可以看到人们在使用它。但我很难弄清楚具体是什么，一个很好的例子是和你谈谈一位科学家使用 AlphaFold，然后加速了药物过程，最终产生了一种我现在可以服用的药物。你最喜欢的例子是什么，关于科学家使用 AlphaFold 做了一些观众可能理解或见过的事情？

Another boon is for researchers who work on neglected diseases that affect primarily the developing parts of the world, things like malaria or Chagas disease or leishmaniasis. These affect hundreds of millions of people around the world. But there's not a lot of money in that if big pharma tried to research that and find cures because they're in the poorer parts of the world. So, the research that goes into that tends to be neglected. There are these amazing non-profit organizations that do the research on that. But they don't have a lot of money or resources, so giving them the structures of the proteins that are involved in, say, malaria virus is a huge boon for them, too, because they can go straight to the drug discovery phase. That was one of the hardest things to figure out as I was doing this research because there's this moment where scientists all around the world have access to AlphaFold. You can see like the map lights up. You can see that people are using it. But I wasn't easily able to figure out what, and a great example would be to talk to you about of a scientist using AlphaFold and then that speeding up a drug process that results in a drug that I could now take. What is your favorite example of a scientist using AlphaFold for something the audience might understand or have seen?

所以，现在有超过 300 万科学家在使用 AlphaFold。我们认为这差不多是全世界每一位生物学家了。

So, over 3 million scientists are now using AlphaFold. We think it's pretty much every biologist in the world at this point.

一家制药公司的一位科学家对我说，从现在起开发的几乎每一种药物都可能在其过程中使用 AlphaFold，这令人震惊且不可思议。但药物发现仍然需要时间。我们仍主要处于基础生物学阶段，了解疾病、我们靶向的蛋白质是什么、那是否是正确的生物学机制。其中一些药物现在正处于临床试验阶段。希望几年后，我们将看到数十种至少部分借助 AlphaFold 开发的药物。

One scientist at a pharma company said to me that almost every drug developed from now on will have probably used AlphaFold in its process, which is mind-blowing and amazing. But it still takes time with drug discovery. We're still mostly in the fundamental biology stage of understanding the disease, what is the protein we're targeting, is that the right biological mechanism. Some of these drugs are now in clinical trials. Hopefully in a few years we'll see dozens of drugs that were partially helped by AlphaFold.

到目前为止，我最喜欢的借助 AlphaFold 取得的突破是核孔复合体。它是体内最大的蛋白质之一，对蛋白质来说非常巨大。它基本上是打开和关闭的通道，让营养物质进出细胞核。就像一个打开和关闭的大甜甜圈环。直到最近我们才知道它的结构，因为它太大太复杂了。在我们发布 AlphaFold 大约 6 个月或一年后，一些团队将其与实验数据结合，最终弄清了这种通道蛋白的美丽形状。这对我来说太神奇了。

My favorite breakthrough so far that happened with the help of AlphaFold is the nuclear pore complex. It's one of the biggest proteins in the body, huge for a protein. It's basically the gateway that opens and closes to let nutrients come in and out of the cell nucleus. It's like a big donut ring that opens and closes. We didn't know until recently what the structure was because it's so big and complicated. About 6 months or a year after we put AlphaFold out, some teams used it along with experimental data to finally work out the beautiful shape of this gateway protein. That was amazing to me.

我们分拆出了一家新公司 Isomorphic Labs，它建立在 AlphaFold 的基础上，并将其作为拼图的一部分，以大幅加速药物发现。平均而言，开发一种药物需要 10 年时间，失败率很高。只有大约 10%的药物能通过所有临床阶段。我们需要大幅改善这一点。方法就是使用计算机模拟方法，AlphaFold 是其中一个组成部分。了解蛋白质的结构只是很小的一部分。你需要大量的化学知识，比如设计什么化合物来与之结合。我们正在构建与更先进 AlphaFold 版本协同工作的相邻系统，并端到端地制造副作用极小且高效药物。我们正在研究大约 18 或 19 个不同的药物项目，涵盖心血管疾病、癌症、免疫学等。最终，这些技术应该能帮助几乎所有治疗领域。

We've spun out a new company, Isomorphic Labs, that builds on AlphaFold and uses it as one piece of the puzzle to massively speed up drug discovery. On average, it takes 10 years to develop a drug, with huge failure rates. Only about 10% of drugs get through all clinical stages. We need to vastly improve that. The way to do that is by using in silico methods, AlphaFold being one component. Knowing the structure of a protein is only one small part. You need a lot of chemistry, like what compound to design to bind to it. We're building adjacent systems that work with more advanced AlphaFold versions, and end-to-end create drugs with minimal side effects and high effectiveness. We're working on about 18 or 19 different drug programs across cardiovascular disease, cancer, immunology, etc. Eventually these technologies should help across almost every therapeutic area.

为此做准备时，我采访了你的诺贝尔奖得主同事 John Jumper。他确实强调这只是药物发现更大问题的一部分。所以，这让我们来到了今天的前沿。现在的前沿是什么？

In prep for this, I did a background interview with your fellow Nobel Prize winner, John Jumper. He really stressed that it's one part of a larger problem of drug discovery. So, that brings us to the cutting edge today. What is the cutting edge now?

我们正在构建许多不同的组件，它们组合在一起。AlphaFold 是一个关键，即蛋白质的结构。但一旦你知道形状，你需要知道蛋白质的哪个部分对其功能重要。要阻断或增强蛋白质，你需要发现一种能附着在正确位置的化合物。你想知道它附着的强度，更重要的是，确保它不附着在其他东西上，否则会导致毒性。利用我们的算法工具，我们可以进行虚拟筛选：AI 系统设计一种化合物，我们预测它与蛋白质表面结合的强度，然后我们可以快速检查它与人体内其他 2 万种蛋白质的附着情况。我们可以修改化合物，使其副作用更少，对靶点的效果更强。这个自我改进过程在计算机模拟中非常快速高效。只有在最后阶段，我们才在湿实验室进行验证。这使我们能够比目前湿实验室的搜索更快、更高效地搜索数千甚至数百万种化合物。

We're building many different components that go together. AlphaFold is one linchpin, the structure of the protein. But once you know the shape, you need to know which part of the protein is important for its function. To block or enhance the protein, you need to discover a chemical compound that attaches to the right place. You want to know how strong it attaches, and more importantly, ensure it doesn't attach to other things, which would cause toxicity. With our algorithmic tools, we can do a virtual screen: an AI system designs a compound, we predict how strongly it binds to the protein surface, and then we can quickly check how it attaches to any of the other 20,000 proteins in the human body. We can modify the compound to have fewer side effects and stronger effect on the target. This self-improvement process is extremely fast and efficient in silico. Only at the final stage do we check in the wet lab. This allows us to search thousands or even millions of compounds more quickly and efficiently than doing the search in the wet lab, which is what is done today.

我喜欢的另一个是 AlphaGenome。我联系了另一位诺贝尔奖得主 Jennifer Doudna 博士，她曾上过我的节目。她给你提了一个问题。所以，我将读出 Doudna 博士的问题。

One of my favorites also is AlphaGenome. I reached out to yet another Nobel Prize winner, Dr. Jennifer Doudna, who I've had on the show. She sent a question for you. So, I'm going to read this question from Dr. Doudna.

她说：「CRISPR，她开创的基因编辑技术，现在可以靶向几乎任何 DNA 序列。但对于大多数遗传病，我们仍然不完全了解 DNA 中的哪些变化实际上导致了问题，尤其是在不编码蛋白质的 98%的基因组中。随着像 AlphaGenome 这样的工具开始解码那 98%，你认为我们距离 AI 能够可靠地指出导致患者疾病的精确基因变化，从而使 CRISPR 这样的技术能够修复它的时刻有多近？」

So, she says, "CRISPR, the gene-editing technology that she pioneered, can now target nearly any DNA sequence. But for most genetic diseases, we still don't fully understand which changes in the DNA are actually driving the problem, especially in the 98% of the genome that doesn't code for proteins. With tools like AlphaGenome starting to decode that 98%, how close do you think we are to the moment where AI can reliably point to the exact genetic change causing a patient's disease so the technologies like CRISPR can fix it?"

是的，这个问题太棒了。实际上，我过去和她讨论过这个问题，这真的很令人兴奋。我认为 AlphaGenome 正是那种技术，它读取很长的基因序列，然后尝试预测：如果你对基因序列中某个特定单字母、单个位置进行了突变，那会是可能导致疾病的有害突变，还是良性的？我们刚刚发布的 AlphaGenome 是世界上预测这方面最好的系统。所以，这正是你想要的。它可能还不够好，但你可以想象未来版本的 AlphaGenome 足够准确，能够真正知道，比如，那个特定的突变与另一个突变结合——这才是难点，如果它们是多基因疾病，由一系列突变引起问题呢？那些更难检测，但实际上非常适合 AI 来帮助解决。然后，也许有一天你可以用像 CRISPR 这样的技术去修复那个突变，从而解决问题。所以，像 AlphaGenome 和 CRISPR 这样的组合可能会非常强大，希望有一天能与 Jennifer 这样的人合作。

Yeah, what an awesome question. So, I've discussed this with her actually in the past and it is really exciting. I think with AlphaGenome, which is exactly that kind of technology, it takes the big long genetic sequences and then it tries to predict if you had made a mutation to this particular single letter, single position in the genetic sequence, will that be a harmful mutation that might cause disease or is it benign? AlphaGenome, which we just released, is the best system in the world for predicting that. So, that's exactly what you want. It's still not probably good enough yet, but you can imagine a future version of AlphaGenome that is accurate enough to really know, like, oh, that particular mutation in combination with this other one—that's the hard part, what if they are multigenic diseases where there are cascades of mutations causing the problem? Those are even harder to detect, but actually perfect for AI to try and help with. Then, you could go in with something like CRISPR maybe one day and fix that mutation and then fix the problem. So, a combination of things like AlphaGenome and CRISPR could be incredibly powerful and hopefully one day will be collaborating with the likes of Jennifer on that.

去年你对《卫报》说了一些我觉得很有趣的话。你说如果按你的意愿，你会把 AI 留在实验室更久。原话是「多做像 AlphaFold 这样的事情，也许能治愈癌症之类的。」从外部看，故事似乎是这样的：你创立 DeepMind，使命是解决智能并用它解决一切。然后你卖给谷歌，正是因为他们允许你以这种方式自由探索科学。很长一段时间里，那是你唯一的焦点。然后 ChatGPT 出现了，谷歌进入红色警戒状态，你成为所有谷歌 AI 的负责人，包括你以前没有花太多时间在的消费产品。从远处看，我觉得这在一定程度上反映了 AI 更大的经历，即过去几年发生的不可思议的变化。在那个变化中，得到了什么，又失去了什么？

Last year you said something to the Guardian that I found really interesting. You said that if I'd had my way, I would have left AI in the lab for longer. And the quote is "done more things like AlphaFold, maybe cured cancer or something like that." From the outside, it looks like the story goes: you found DeepMind with the mission to solve intelligence and use it to solve everything else. And then you sell to Google specifically because they will allow the freedom to explore science in this way. And for a long time that's your exclusive focus. And then ChatGPT comes out, Google goes code red, and you become the head of all Google AI including the consumer products that you weren't spending as much time on before. And it feels to me like watching that from afar, it mirrors somewhat the larger experience of AI which is just this incredible change in the last couple of years. What was gained and what was lost in that change?

是的，我认为你描述得完全正确，从内部来看感觉也是如此。对我来说，正如我之前提到的，AI 的最佳用例是改善人类健康和加速科学发现。事实上，我最初进入 AI 领域是因为我对世界上所有重大问题感兴趣——现实的本质、意识的本质等等——我觉得我们需要一个工具来帮助即使是最优秀的科学家理解大量的数据和信息，并从中发现洞见。而这正在发生，这太棒了，显然 AlphaFold 是我们第一个，也是迄今为止最好的体现。我一直想着这个以及其他许多类似的问题。所以，我认为那会很棒，考虑到 AGI 的重要性以及它作为一项变革性技术——可能是人类历史上最具变革性的技术——那么我认为最好以非常谨慎、精确、深思熟虑和严谨的科学方法来处理我们现在所处的构建 AGI 的最后阶段，在我理想的世界里，像 CERN 那样的合作努力，确保每一步我们都理解每一步，直到实现构建 AGI 的最终目标。对于这样的技术，这似乎是最合理的。当然，你不必等待，所以那可能需要更长的时间，也许十年甚至二十年，但我认为考虑到我们正在处理的事情的规模，这是有道理的。然后我的另一个想法是：我们不必等到 AGI 到来才开始获得 AI 的好处。我们可以使用更专门的系统，这些系统可能利用我们为 AGI 开发的通用技术、通用算法，但它们本身不是通用智能。它们是窄 AI，如果你愿意这么叫的话，比如 AlphaFold，它只做特定目的，而且只做那个目的。我们可以——而且我们仍然在，我仍然在做——在谨慎科学地构建 AGI 的同时，创造许多类型的 AlphaFold 和 Isomorphic，然后人类可以从其成果中受益，比如治愈癌症，或者新的能源或新材料。所以，我觉得，也许从二三十年前我刚开始这一切的时候看，那是我认为的理想方式。现在，事情并没有那样发生，因为技术是不可预测的。事实上，事实证明像语言这样的东西比我们所有人预期的要容易得多，甚至包括我们这些对整体技术明显乐观的人。最终我们会攻克语言，但现在想起来很有趣，但语言、概念和抽象——当前模型，如 Gemini 这样的基础模型做得非常好——我们以为可能还需要一两个或三个突破才能达到。但事实证明，Transformer（我的谷歌同事发明的）以及一些强化学习就足以攻克语言之类的东西。我们和其他领先的实验室在某种程度上一直在摆弄它，但当然，随着 ChatGPT 的出现——公平地说，OpenAI——他们扩展了它，然后把它推出来，我认为甚至他们自己都说那是一种科学实验。

Yeah, I think that's exactly right what you described, it's sort of how it felt from the inside too. And for me, as I mentioned earlier, the best use case of AI was to improve human health and accelerate scientific discovery. In fact, I got into AI in the first place because I was interested in all the big questions in the world—the nature of reality, nature of consciousness, these kinds of things—and I felt we needed a tool to help even the best scientist make sense of the amount of data and information out there and find insights in that. And that's happening, which is amazing, and obviously AlphaFold was our first and, so far, best expression of that. And I always had that on my mind and many other problems like that. So, it would have been great, I think, and given how important AGI is and how transformative a technology it is—maybe the most transformative one in human history—then I thought it would be best to approach these kinds of latter stages of building it, which we're in now, using the scientific method very carefully, very precisely, very thoughtfully and rigorously with all the best scientists, in my ideal world collaborating in a CERN-like effort on making sure each step we understood each step as we got to the final goal of building AGI. That would make the most sense with a technology like this. And of course you don't have to wait, so that might take a lot longer, maybe a decade or even two decades longer, but I think that would make sense given the enormity of what we're dealing with. And then my other idea was: we don't have to wait till AGI arrives to start getting the benefits of AI. We could use more specialized systems that maybe make use of the general technologies, the general algorithms we're developing for AGI, but are not in themselves general intelligences. They are narrow AIs, if you want to call them, like AlphaFold which does a specific purpose and only that purpose. And we could—and we still are, I'm still doing this—create many types of AlphaFolds and Isomorphics while we're building AGI in this careful scientific way, and then humanity could benefit from the proceeds of that, like cures for cancer or maybe new energy sources or new materials. So, I felt that would be, maybe looking at this from 20-30 years ago when I started out on all of this, the ideal way for it to play out in my opinion. Now, it didn't happen like that because technology is unpredictable. In fact, it turns out that things like language were a lot easier than we were all expecting, even those of us who were obviously optimists about the whole technology. Eventually we'll crack language, but it seems funny to think of it now, but language and concepts and abstractions—things that the current models, foundation models like Gemini, do incredibly well—we thought that maybe there would be one or two or three more breakthroughs needed before we could get there. But it turned out transformers, which my Google colleagues invented, and some reinforcement learning as well on top, was enough to crack things like language. And we were sort of playing around with that with the other leading labs, but of course with ChatGPT—and fair play to OpenAI—they scaled it and then they put it out there, and I think even they say it was a sort of science experiment.

他们没意识到它会这么火，我想我们谁也没料到，当时我们的系统水平差不多。因为当你构建这项技术时，你离它太近了，非常清楚它做不到的事、它的缺陷，没意识到外面的人其实会觉得它有用，尽管它会胡编乱造，还有其他我们现在仍在努力改进的问题，虽然还没完全解决，但仍有有趣的用例，比如总结、头脑风暴之类的，现在大家都用聊天机器人做这些事。不好的一面是，我们陷入了激烈的商业压力竞赛，每个人都被卷进去了。除此之外还有地缘政治问题，比如中美竞赛等等。所以有多个层面的压力要求快速行动。好处当然是进步更快，现在进展快如闪电，这对所有好的用例都有利。第二个好处是，所有观众，每个人，都能用上最前沿的 AI 技术，可能只比实验室里落后三到六个月。这很惊人。这也很好，因为我觉得它让每个人感受到 AI 的民主化。它让每个人体验到与前沿 AI 互动是什么感觉，它能做什么、不能做什么。我认为这对社会有好处，让它开始适应这项技术即将带来的巨大变化。所以，我们逐步体验它可能比突然冲击更好。比如今天没有 AGI，某天突然有了 AGI，那可能不好，尽管我认为本来有很多种推出方式。最后一点好处是，你无法真正完全了解你的系统，除非经过数百万人的压力测试。所以无论你的测试多好，内部测试当然有数百万聪明人尝试各种东西，然后你看到什么浮出水面，或者你得到的反馈，对构建更稳健、更好的系统非常重要。所以我认为这样发展有利有弊。这不是我多年前梦想的方式，那时我们本可以更哲学地思考这件事。所以我们必须面对现实的世界，并尽力而为。我们通过推进前沿来做到这一点，同时也尽可能负责任地部署像 Gemini 和 AlphaFold 这样强大的技术。

They didn't realize it would go so viral and I think none of us did and we had sort of fairly equivalent systems at the time because I think when you're building that technology, you are so close to it, you're very aware of the things it can't do, the flaws it has and you don't realize that actually people out there would find use even though it was hallucinating and doing other things that we're obviously all still trying to improve on now, still not completely fixed, but there's still interesting use cases like summarizing things or brainstorming things like that that people use, everyone uses chatbots for today. Now, the downside of it is that we're in this sort of ferocious commercial pressure race that everyone's sort of locked into currently. And then on top of that there are geopolitical issues like the US-China race and so on. So, there are sort of multiple levels of pressure to move fast. So, the benefit of that is of course you get faster progress obviously. So, the progress is just like at lightning speed these days. So that's good for all the good use cases. The second benefit is that all the viewers out there, everyone, you're all getting to use the most cutting-edge AI technology perhaps only three to six months behind what is actually in the labs. So, that's kind of mind-blowing. It's also great because I think it gives everyone a feeling for it's democratizing AI. It's giving everyone a feeling for what it's like to interact with cutting-edge AI and what it can do and what it can't do. And I think that's good for society to start normalizing itself to what is going to be an enormous change with this technology coming. So, it's probably better that we get to sample that in incremental steps rather than it's just a shock to the system. Here's no AGI and then here's AGI one day. Probably that's not good although I think there could have been many ways it could have rolled out. And then the final thing on the benefit side is that you can't really fully understand your systems until they're stress tested by millions of people. So, it doesn't matter how good your testing is, your in-house testing obviously millions of smart people trying out things and then you seeing what bubbles to the top or the feedback you get is really important for building more robust systems and better systems. So, I think there are positives and negatives about how it's gone. It's not the way I dreamed about years ago where we would be sort of contemplating this philosophically and so, we have to deal with the world as we find it and make the best of that. And we try to do that by advancing the frontier, but also trying to be as responsible as we can with doing that as we deploy these very powerful technologies like Gemini and AlphaFold.

与此同时还有另一个故事，我想回到你的担忧以及你如何权衡这些担忧和代价。为了理解这一点，我认为我们需要讲一个关于 AI 非常有创造力、出乎意料地有创造力的故事。这个故事开始于——让我找到我的积木块。故事从这里开始。让我们回到 2016 年 3 月 10 日。一位非常著名的围棋选手坐下来与你们设计的系统对弈。当时计算机已经在各种游戏中击败人类，但围棋非常有趣，因为围棋的潜在走法比宇宙中的原子还多。他们来回对弈。然后你的系统走出了一步令人震惊的棋，因为人类几乎不可能想出那样的走法——第 37 手。

There's another story happening at the same time as this and I want to get back to your concerns and how you weight those concerns and the cost. In order to understand that, I think we need to tell a story about AI being very creative, unexpectedly creative. And that story begins let me find my Jenga block. That story begins here. So, let's go back to March 10, 2016. There's a very famous Go player that sits down to play against a system that you designed. And at this point computers have beat humans at all kinds of games, but Go is really interesting because there are more potential moves in Go than atoms in the universe. They go back and forth, they're playing. And then your system makes a move that is so surprising because it is incredibly unlikely that a human would figure out a move like that, move 37.

是的。

Yes.

你看到李世石坐在那里，脸上满是震惊。他双手抱头。那一刻，我认为像你这样的人预见到了 AI 系统的创造力，这与我们之前讨论的系统截然不同。有一种类型是给大量数据，要求做出新预测。我知道这比这个简化复杂得多。但还有一种类型是不给数据，只给规则。

And you see Lee Sedol sitting there, he's just got this shock on his face. He's got his head in his hands like this. And it really was this moment where I think people like yourself saw ahead to the creativity that we would find in AI systems that are very different than the systems that we've talked about so far. So, there's a category where you're giving a huge amount of data and you're asking to make new predictions. And I understand this is much more complicated than this oversimplification. But then there's a category where you're not giving data, you're giving rules.

嗯。

Mhm.

比如数学、物理或围棋这样的游戏。它有着不可思议的创造力机会。

Like with math or physics or games like Go. And it has this incredible opportunity for creativity.

是的。

Yeah.

那一刻你在哪里？你看到了什么样的未来？

Where were you when that moment happened? And what future did you see ahead?

是的，你描述的是一个不可思议的时刻，实际上差不多正好是十年前，感觉像是一个世纪前，但我认为在很多方面，那是现代 AI 时代的黎明。因为在那之前，有很多 AI 程序能在游戏中成为世界冠军，比如国际象棋，但它们是用所谓的专家系统做的。这些系统由一群聪明的程序员和一群聪明的国际象棋特级大师合作，试图将特级大师的知识提炼成一套规则和系统，一种蛮力系统，会使用大量算力，比如 IBM 用深蓝击败加里·卡斯帕罗夫那样。他们把象棋专家给的规则封装起来，然后系统机械地执行这些规则和启发式方法，进行数百万次的走法搜索，然后根据这些启发式方法找出最佳走法。但对我来说，90 年代看到这个并不满意。我当时在读本科。我觉得那不是真正的 AI，因为那个系统，比如深蓝，它在国际象棋上是世界冠军水平，但其他什么都不会。不仅不会语言、机器人等，连更简单的井字棋都不会，对吧？所以智能的定义显然有问题。没有人类特级大师学不会井字棋，那说不通，因为井字棋严格更简单。所以它的泛化能力有问题，而且它没有学习，只是被给了答案，对吧？如果你问深蓝这样的系统，智能在哪里？它不在系统里，而在国际象棋特级大师和程序员的头脑里。他们解决了国际象棋问题，然后实现了解决方案。程序只是机械地执行解决方案。而围棋，正如你所说，是游戏的最后前沿。

Yeah, it was an incredible moment that you're describing and it's actually almost exactly 10 years ago now, which feels like a century ago actually, but I think in many ways it was the dawn of the modern AI era because until that point there were many AI programs that could be world champions at games, things like chess, but they were done with what's called expert systems. So, they were systems where a team of smart programmers with a team of smart, in that case chess grandmasters, came together, tried to distill the knowledge the chess grandmasters have into a set of rules and system, kind of a brute force system, that would use a lot of compute like on a supercomputer like IBM did with Deep Blue to beat Garry Kasparov, and they would in sort of encapsulate the rules they were given by the chess experts, and then the system would sort of dumbly execute those rules and heuristics and do millions and millions of searching of moves, and then try and work out against those heuristics which is the best one to do. Now, the thing with that is, for me that was not satisfactory when I saw that in the '90s. I was doing my undergrad at the time. I didn't feel like that was proper AI because that system, let's take Deep Blue, okay, it's world champion level at chess, but it can't do anything else. Not only can't it do language and robotics or any of those kind of things, it can't even play a strictly simpler game like tic-tac-toe, right? So, something's obviously not quite right about the definition of intelligence, right? In the sense that no human, you could imagine a human grandmaster not being able to learn how to play tic-tac-toe. It would make no sense because it's strictly simpler. So, there's something sort of wrong about its generalization capability and the fact that it didn't learn. It was just given the answer, right? So, if you could ask for something like Deep Blue, where did the intelligence reside of the system? Well, it wasn't in the system, it was in the minds of the chess grandmasters and the programmers. They solved the problem of chess and then implemented the solution. The program just dumbly executed the solution. Now, Go, as you mentioned, is the sort of final frontier for games.

围棋是人类发明的最复杂的游戏。它也是最古老的游戏，所以在很多方面都很了不起。而且它也非常优美。在亚洲，中国、日本和韩国下围棋，它占据了相当于国际象棋的智力阶层。但围棋是一种更直觉性的游戏，几乎是一种艺术性的游戏。你下出看起来很美的棋形，结果证明它们真的很强，这就是为什么围棋带有一些神秘色彩。顶尖棋手会说围棋蕴含着宇宙的奥秘。我认为古代中国人就是这样想的。还有它纯粹的复杂性：可能的棋盘位置有 10 的 170 次方，比宇宙中的原子还多。这意味着你无法像我们处理国际象棋那样用暴力计算来解决它。此外，因为围棋如此直觉和深奥，没有可以轻易封装成机器遵循的规则。当你问一位围棋大师时，不像国际象棋大师，他们会说：「你为什么下在那里？」他们会说：「感觉对了。」国际象棋棋手永远不会这么说；他们会说：「我这么做是因为我在计算这个、这个，」然后告诉你计算过程。那种直觉显然很难封装进系统。你无法直接编程实现。所以，它是我们早期在 DeepMind 开创的新技术——深度强化学习——的完美试验场。你能构建从自身经验中直接学习的系统吗？

It's the most complex game humans have ever invented. It's also the oldest game, so it's amazing in many ways. And it's also very beautiful. In Asia, where they play in China, Japan, and Korea, it occupies the intellectual echelon instead of chess. But it's a much more intuitive game, sort of artistic game almost. You play patterns that look beautiful and they turn out to be really strong, which is why the game has a mystical element. Top Go players would say it encapsulates the mysteries of the universe. I think that's how the ancient Chinese thought about it. And also its raw complexity: more possible board positions, 10 to the power 170, than there are atoms in the universe. That means there's no way you can brute force it like we did with chess. Furthermore, because the game is so intuitive and esoteric, there aren't rules you can easily encapsulate for a machine to follow. When you talk to a Go master, unlike a chess master, they'll say things like, "Why did you play there?" They'll say, "It felt right." A chess player would never say that; they would say, "I did it because I'm calculating this, this," and then tell you the calculation. That intuitive feeling is obviously very hard to encapsulate in a system. You can't really program that directly. So, it's the perfect proving ground for the new techniques we were pioneering in the early days of DeepMind: deep reinforcement learning. Can you build systems that learn from themselves directly from experience?

以 AlphaGo 为例，它首先查看互联网上所有人类下过的棋局，学习人类会下的棋步，然后我们叠加了蒙特卡洛树搜索，让它能够发现围棋知识树的新分支。从人类已知的开始，然后超越它们。这正是我们所希望的。那场比赛最终被全球 2 亿人观看，令人惊叹的是，我们不仅以 4 比 1 赢得了比赛，而且在第二局中，它下出了著名的第 37 手，一个创造性的棋步，在棋盘第五线，开局阶段。在围棋中这是大忌；如果围棋大师教你，他们会打你的手腕，因为那被认为是坏棋。但不仅是一手好棋，它最终为 AlphaGo 赢得了比赛。100 手、200 手之后，它正好在正确的位置，仿佛有先见之明地放了一颗棋子。所以这是关键的一手，不仅令人惊讶，而且对后来至关重要。显然它改变了所有围棋棋手下棋的方式，但对我来说，这是我等待已久的时刻：构建一个系统，能够实现其他系统无法做到的事情，游戏 AI 的珠穆朗玛峰，击败围棋世界冠军的最后前沿。而且，不仅是赢了，而是以第 37 手这样的创造性新想法获胜。这对我来说是信号，表明我们可以转向像 AlphaFold 这样的科学问题了。

In the case of AlphaGo, it started by looking at all the games on the internet that humans have played and learning the types of moves humans would do, but then we overlaid it with a Monte Carlo tree search that allowed it to discover new branches of the tree of knowledge in Go. Starting with what humans knew and then going beyond that. That's what we hoped would happen. The amazing thing about that match, which ended up being watched by 200 million people around the world, was that not only did we win the match 4-1, but in game two, it played this famous move 37, a creative move on the fifth line of the board early in the game. It's a big no-no to do that in Go; if you were taught by a Go master, they would slap your wrist for playing there because it's regarded as a bad move. But not only was it a great move, it ended up winning the game for AlphaGo. 100 moves, 200 moves later, it was in the right place, as if it presciently put the stone there. So it was a critical move, not only surprising but decisive for later. Obviously it changed the way all Go players play Go, but for me it was the moment I'd been waiting for: building a system that could achieve something no other system could, the Mount Everest of games AI, the final frontier of beating the Go world champion. But also, it was how it won, with these creative new ideas like move 37. That for me was the signal that we were ready to turn to scientific problems like AlphaFold.

为什么这个想了解未来的听众需要理解第 37 手和围棋发生的事情很重要，因为这意味着如果 DeepMind 能构建一个做到这一点的系统，它也许也能构建一个能玩任何游戏的系统。它也许还能构建系统，在现实世界问题中找出最佳解决方案，比如量子计算、核聚变、矩阵乘法、芯片设计等等。

The reason why it's important that this audience that wants to understand the future understand what happened with move 37 and Go is because the implication is if DeepMind can build a system that can do that, it can also perhaps build a system that can play any game. It can also perhaps build systems that can figure out in real world problems what is the best solution in quantum computing, nuclear fusion, matrix multiplication, chip design, etc.

是的。

Yes.

你能谈谈前沿进展吗？选一个系统。那个令人惊讶的创造性元素，相当于第 37 手的是什么？

Could you tell me about the cutting edge here? Pick one of these systems. What is the move 37 of the surprising creative element going on?

我认为 AlphaZero 非常值得讨论，它是 AlphaGo 的进化版。在我们达到围棋巅峰并展示它能提出像第 37 手这样的新想法后，我们进一步将其泛化为一个名为 AlphaZero 的系统，我认为它今天也会变得非常重要。对于 AlphaGo，我们从互联网上能找到的所有人类棋局开始，并且系统中内置了一些围棋特有的东西，比如棋盘的对称性。我们想完全摆脱所有这些假设，从头开始，就好像程序和算法对要解决的问题一无所知。这就是 AlphaZero 中「零」的含义：去除任何人类手工知识，无论是数据还是启发式方法。AlphaZero 像白板一样开始。显然它有一个学习系统，一个神经网络，我们设置了参数，但没有给它任何关于围棋或其他游戏的领域特定知识。然后我们测试 AlphaZero：首先，它能从零开始学习围棋并击败 AlphaGo 吗？我们做到了。它需要 17 次迭代。AlphaZero 开始时是随机的，只知道游戏规则，随机下棋。显然很糟糕。它通过与自己下 10 万局棋来创建自己的数据集。然后它可以看哪些棋步赢了或输了。即使它基本上随机下，也会有一些棋步比其他稍好。所以它用这 10 万局棋，用新数据训练自己的第二个版本。第二个版本比第一个稍好。现在它不再是随机的了，但还不强，下出还可以的棋步。然后那些还可以的棋步变得更好。所以版本二训练成版本三、版本四，以此类推。

I think AlphaZero is very interesting to talk about, which was the evolution of AlphaGo. After we reached the pinnacle of Go and showed it could come up with new ideas like move 37, we generalized it further to a system called AlphaZero, which I think will turn out to be very important for today as well. With AlphaGo, we started with all the human games we could find on the internet, and there were a few other things specific to Go built into the system, like the symmetry of the board. We wanted to get rid of all those assumptions completely and start from scratch, as if the program and algorithm didn't know anything about what it was trying to do. That's what the "zero" refers to in AlphaZero: removing any human-crafted knowledge, both in data and heuristics. AlphaZero starts like tabula rasa. Obviously it has a learning system, a neural network, we set up the parameters, but we didn't give it any domain-specific knowledge about Go or any other game. Then we tested AlphaZero: first, could it learn Go from scratch and beat AlphaGo? And we managed to do that. It takes 17 iterations of the program. AlphaZero starts off random, only has the rules of the game, plays randomly. Obviously it's terrible. It creates its own dataset by playing 100,000 games against itself. Then it can see which moves won or lost. Even though it's playing more or less randomly, there will be some moves slightly better than others. So it takes the 100,000 games, we train a new version of itself on version two with that new data. Version two is slightly better than version one. So now it's not random anymore, but not great, playing okay moves. Then those okay moves end up being better. So version two gets trained to version three, version four, and so on.

所以，每次新系统与旧系统对弈，看看它是否显著更好。结果发现，至少在围棋和国际象棋这类游戏中，大约 16 到 17 代就足以从随机水平达到超越世界冠军的水平。至少在国际象棋中，我实际上曾亲眼目睹过，因为我很着迷，我自己也下棋。它从早上开始是随机的，到午餐时间我还能勉强与它竞争，到下午茶时间它已经超越所有特级大师，到晚餐时间它已经超越世界冠军。你亲眼看到了从零开始的整个进化过程。而且它下出了有趣的新棋步，即使是像 Stockfish 这样更依赖专家系统和暴力计算的国际象棋计算机，也没有发现这些新型棋步。所以，AlphaZero 是 AlphaGo 思想的完全泛化。

And so, each time that new system gets played against the old system and sees if it is significantly better or not. And it turns out that at least in Go and chess and things like that, around 16 or 17 generations of that is enough to go from random to better than world champion. And at least in the case of chess, which I actually once watched live happen because I was fascinated by it, I was even playing chess myself. It starts in the morning random, then by lunchtime I could still just about compete with it myself, and then by tea time it's better than all grandmasters, and then by dinner time it's better than the world champion. And you've just seen the entire evolution of that from scratch. And also it's playing interesting new chess that even chess computers like Stockfish, with the more expert system brute force ones, haven't discovered those types of new moves. So, AlphaZero was the full generalization of the AlphaGo ideas.

有趣的是，我认为我们现在需要将这些思想重新应用到我们的基础模型上，比如新的 Gemini 这类东西，你可以把它们看作是对一切事物——语言、我们周围的世界——的通用模型，而不仅仅是像围棋这样的游戏。但我们仍然需要在这些模型之上进行搜索、思考和推理的能力。我们有时称这些为世界模型。我认为如何做到这一点还没有完全解决。重新引入一些 AlphaGo 的思想，但现在不是仅仅应用于一个狭窄的游戏，而是应用于整个世界，也许还有科学的部分领域，比如材料设计、芯片设计、量子计算机。所有这些酷炫的项目，太多了，我只想看到所有这些积木。我简直不敢相信我们真的在做所有这些事情，但这是真的。这有点像梦想：我热爱每一个科学分支，我可以在所有这些不同的科学领域尽情投入，因为 AI 是一个如此通用的工具，它确实可以在所有这些领域产生巨大影响。

And interestingly, I think we need these types of ideas back here now with our foundation models, the new Gemini and these kinds of things, which you can think of as generalized models of everything, language, the world around us, not just a game like Go. But we still need this ability to search and think and reason on top of those models. And sometimes we call those world models. And I think that still hasn't fully been cracked yet how to do that. Bringing back some of these AlphaGo ideas, but now instead of just a narrow game applying it to that, but to the whole world, and maybe, interestingly, parts of science too, like material design, and things like chip design, and quantum computers. All of these cool projects, there are so many, I just want to see all these bricks. I can't believe we're actually working on all these things, but it's true. And this is sort of the dream: I love every branch of science, and I get to indulge myself in all these different areas of science because AI is such a general tool, it can really make a huge difference to all these areas.

所以，我举的一个例子就是设计新材料。如果你想要一种具有特殊性质的材料，我们能否超越材料科学目前已知的范围？我认为类似 AlphaGo 的过程在那里会非常有用。而相当于第 37 手棋的，就是 AlphaTensor 找到一种新算法，使矩阵乘法变得更好。

So, maybe one example I give is just designing new materials. If you want a material with a special type of property, can we go beyond what is currently known in material science? And I think AlphaGo-like processes could be very useful there. And the equivalent of a move 37 would be like AlphaTensor finding a new algorithm that makes matrix multiplication better.

更快。

Faster.

完全正确。所以，你可以将其应用于算法空间，这非常令人兴奋，因为算法本身变得更快，所以存在某种循环改进。是的，AlphaTensor，仅仅是改进矩阵乘法，而矩阵乘法是所有神经网络的基础。事实证明一切都是矩阵乘法。如果你让它快 5%，考虑到训练花费的数百亿美元，那将节省巨大成本。所以这些都是很好的例子。而且我认为我们还处于早期阶段。还有像芯片设计中的布线问题，使其尽可能高效。这是一种 NP 难问题，类似于旅行商问题：连接所有这些组件的最短距离是多少。像 AlphaChip 这样的程序非常擅长处理这个问题，在某些情况下甚至比人类芯片设计师更好。所以，我认为我们才刚刚触及表面，未来几年，当今更通用的系统与 AlphaGo 和 AlphaZero 的这些思想相结合，将实现更多可能。这两类——从 AlphaFold 开始的故事，从 AlphaGo 开始的故事——是让我感到非常乐观的 AI 类型。

Exactly. Exactly. So, you can apply it in algorithmic space, which is quite exciting because then the algorithm itself gets faster, so there's some circular improvement there. And yes, AlphaTensor, just making matrix multiplication, which is the basis of all neural networks. It turns out everything is matrix multiplication. If you just make that 5% faster, that's a huge cost saving given the tens of billions being spent on training. And so these are good examples of ideas. And I think we're still early. Also like things like the design of chips on a die, making it as efficient as possible the routing. It's a kind of NP-hard problem, like the traveling salesman problem: what's the shortest distance you can wire up all of these things. And programs like AlphaChip are really good, better in some cases than human chip designers at dealing with that. So, I think we're just scratching the surface of what's going to be possible in the next few years with today's more general systems combined with these types of ideas from AlphaGo and AlphaZero. These two categories, the story that starts with AlphaFold, the story that starts with AlphaGo, these are the kinds of AI that make me feel really optimistic.

我也认为，真正乐观——而且你在公开场合经常这样做，我很欣赏——就是充分思考事情可能出错的方式以及我们可以做些什么来防止它。

I also think that being really optimistic, and you do this a lot in public, which I appreciate, is fully thinking through the ways in which something can go wrong and what we can do to prevent that.

是的。

Yeah.

所以，我想插入另一个话题。当然。我们在做什么？这个。是的。我提起这个游戏的原因是因为这是一个实时战争游戏。在视频中，这个系统完全碾压人类，你可以看到工程师们为他们的系统胜利而欢呼。但当然，作为一个没有构建这个系统的人，我在想，如果那是真的呢？我们现在正在谈论，关于军队和政府使用 AI 的辩论是一个巨大的话题。我希望这个对话能持续 10 年。我希望它能在那段时间内有用。所以，我不想谈论具体的公司、具体的服务条款。我也认为人们在某种程度上只见树木不见森林。因为从更大的图景来看，政府将会使用 AI。所以，我想从你——构建这些系统的人——那里知道，如果你能挥动魔杖，你希望他们用它来做什么？

So, I want to insert one other in here. Sure. What are we doing? This one. Yes. And the reason why I bring up this game is this is a real-time war game. And in the videos where this system is absolutely crushing humans, you can see the engineers cheering for the victory of their system. But of course, as someone who didn't build the system, I'm thinking to myself, what if that's real? And we're speaking right now during a time when the debate about militaries and governments using AI is a huge topic of conversation. I want this conversation to last for 10 years. I want it to be useful for that long. So, I don't want to talk about specific companies, specific terms of service. I also think people are in some way missing the forest for the trees here. Because bigger picture, governments are going to use AI. And so, what I want to know from you as someone building these systems is if you could wave your magic wand, what would you hope that they use it for?

嗯，我认为政府应该使用 AI，我们希望支持所有民主选举的政府。我希望看到他们将其用于改善公共卫生、教育等领域。我的意思是，所有这些都需要重新思考。效率的提升和我们能用它做的好事，政府可以为公民做到，这将是不可思议的。我认为一些国家正在这样做，比如新加坡和阿联酋，它们正在倾向于这类用例。我希望看到它被用于能源领域，比如优化电网。我们在数据中心这样做了，节省了 30%用于冷却系统的能源。我认为在这些领域大规模应用 AI 将带来巨大的社会收益。所以，这是我一直在思考的，并希望政府能够采纳和使用，我们愿意支持这一切。当然，当前世界的地缘政治非常复杂，而这些是双重用途技术。我担心 AI 可能出错的一些用例。从更大的图景来看，正如你所说，我认为有时人们会陷入细节，但实际上，大图景中有两件事需要担心。

Well, look, I think governments should be using AI, and we want to support all democratically elected governments. And I think the things I would love to see them use it for and what we're trying to build our systems to be good for is things like improving public health, education. I mean, all of these things need to be rethought. The efficiency gains and the amount of good we can do with it, governments could do with it for their citizens could be incredible. And I think some countries are doing it like Singapore and UAE, I think are leaning into these types of use cases. I would love to see it being used for things like energy, like optimizing energy grids. We did that with our data centers and saved 30% of the energy used for the cooling systems. I think there's enormous societal gain from applying AI at scale to these types of areas. So, that's what I've always thought about and hope that governments will pick up and use for, and we want to support all of that. Of course, the geopolitics of the world is very complicated right now, and these are dual-purpose technologies. And I worry about a couple of use case things that can go wrong with AI. In the bigger picture, as you say, I think sometimes people get bogged down in the details, but actually the big picture, there are two things to worry about.

一是恶意行为者，无论是个人还是国家，将我们试图用于善举的技术——比如治疗疾病、推进材料科学和能源——重新用于有害目的，无论是有意还是无意。我担心的第二个方面是，随着系统变得更强大，AI 本身可能会失控。这不是今天的系统，但可能在接下来的两到四年内，尤其是当我们进入智能体时代时。所谓智能体，是指能够自主完成整个任务的系统。我们需要它们，因为它们会非常有用，比如作为助手，但它们也会变得越来越有能力和自主。因此，作为前沿实验室之一，我们必须确保设置好护栏，让这些系统完全按照指令行事，目标被足够清晰地指定，并且无法绕过或意外突破这些护栏。考虑到这些系统最终会变得多么强大和智能，这是一个极其困难的技术挑战。我倾向于担心这些中期问题——尽管三四年算不上中期——但我认为人们目前对它们关注不够。如果我们想以对人类有益的方式度过 AGI 时刻，这些将是我们必须应对的最大问题。

One is bad actors, whether individuals or nation states, repurposing these technologies we're trying to build for good—like curing diseases, advancing material science and energy—for harmful ends, either inadvertently or intentionally. The second branch I worry about is the AI itself going rogue as systems become more powerful. That's not today's systems, but maybe in the next two to four years, especially as we enter the agentic era. By agents, I mean systems capable of completing entire tasks on their own. We want them because they'll be very useful, like as an assistant, but they'll also be increasingly capable and autonomous. So, as one of the frontier labs, we must ensure guardrails are in place so that these systems do exactly what they've been told, with goals specified clearly enough, and no way of circumventing or accidentally breaching those guardrails. That's an incredibly hard technical challenge given how powerful and smart these systems will eventually become. I tend to worry about these medium-term issues—though three or four years isn't really medium term—but I think people aren't paying enough attention to them. They will be the biggest issues we have to contend with if we're to get through the AGI moment in a way beneficial for humanity.

我带来的最大问题之一是：下次我看到头条新闻时，如何权衡我们在未来 30 年都会有的担忧？人们过度担心什么，又对什么担心不足？

One of the biggest questions I came in with was: next time I read a headline, how do I weight the concerns we'll all have over the next 30 years? What are people worrying too much about, and what aren't they worrying enough about?

我刚才提到的两件事，是普通人甚至一些专家和科学家可能担心不足的。它们是对社会影响的关键问题。还有其他短期担忧，比如深度伪造和 misinformation。我们开发了一个名为 SynthID 的系统，这是一种 AI 水印，可以对任何生成的图像进行数字水印。所有谷歌技术，如 VO 等，都内置了这种水印技术，以便我们可以检测并向用户或政府标记伪造内容。我主张所有从事生成式 AI 的公司都应该构建类似的技术，这样至少可以检测出来。这将变得越来越重要，但与 AGI 本身变得非常强大以及我们如何确保设置护栏这些更大的问题相比，它就显得微不足道了。我们需要在这方面投入更多的研究和努力，我希望能看到领先实验室、AI 安全研究所和学术界之间的国际合作，共同应对下一步，因为创造这样的技术是前所未有的。

The two things I just mentioned are what the average person—and even some experts and scientists—might not be worrying enough about. They are the key societal-affecting issues. There are other immediate-term worries like deepfakes and misinformation. We work on a system called SynthID, an AI watermark that digitally watermarks any generated image. All Google technologies, like VO and others, have this watermarking technology so we can detect and flag fakes to users or governments. I advocate that all companies working on generative AI should build in similar technology so that at least it can be detected. That will become increasingly important, but it pales compared to the bigger issues around AGI itself becoming very capable and how we ensure guardrails are in place. A lot more research and effort needs to go into that, and I'd love to see international cooperation among leading labs, AI safety institutes, and academia to navigate this next step, because creating such technology is unprecedented.

如果我们推演下去，极限在哪里？你认为 AI 不能做而人类能做的事情是什么？你曾称这是你人生的核心问题。

If we play this out, what's the limit? What are the things you think AI cannot do that humans can? You've called this the central question of your life.

是的，这与我的一些英雄如艾伦·图灵的思考有关。他描述了图灵机——一种理论构造，所有现代计算机基本上都是图灵机，能够计算任何可计算的东西。任何可以描述为算法的东西，这台机器都能计算。我认为我们正在构建的系统是近似的图灵机，许多神经科学家（包括我）认为大脑可能被建模为近似的图灵机。但其他人，比如我的朋友罗杰·彭罗斯，认为大脑中可能存在量子效应。我们有过友好的辩论，但到目前为止，神经科学还没有在大脑中发现任何量子效应。看起来大脑的大部分活动是经典计算。因此，AI 系统最终能做什么和模仿什么，其极限并不明确。这是一个实证问题。关于意识的问题定义不清，但我们凭直觉知道它是什么。构建智能人工制品的旅程将提供与人类思维的对照研究，我们将看到思维的独特之处。我持开放态度；可能存在独特的东西和人类之间独特的联系，这些永远不会被 AI 复制。但许多目前无法企及的事情，比如长期规划、推理和某些形式的创造力，我认为 AI 最终将能够做到。我想诚实地说：我正在做人类历史上一直在做的事情——试图找出我们为什么特殊。

It is, and it's related to the thinking of some of my heroes like Alan Turing. He described Turing machines—theoretical constructs that all modern computers are basically Turing machines able to compute anything computable. Anything that can be described as an algorithm, this machine can compute. I think the systems we're building are approximate Turing machines, and many neuroscientists, including me, think the brain might be modeled as an approximate Turing machine. But others, like my friend Roger Penrose, believe there might be quantum effects in the brain. We've had good-natured debates, but so far neuroscience hasn't found any quantum effects in the brain. It looks like most brain activity is classical computation. So it's not clear what the limit would be for what an AI system could do and mimic. That's an empirical question. Questions around consciousness are not well-defined, but we intuit what it is. This journey of building an intelligent artifact will provide a controlled study comparison to the human mind, and we'll see what's unique about the mind. I'm open-minded; there could be unique things and unique connections between humans that will never be replicated by AI. But many things currently out of reach, like long-term planning, reasoning, and some forms of creativity, I think AI will eventually be able to do. I want to be honest: I am doing exactly what humans have done throughout history—trying to find why we are special.

我们必须是宇宙的中心。哦等等，我们不是。我们必须是情感共鸣的物种。哦等等，大象也有葬礼。哦，我们一定是能创造艺术的物种。哦等等，Gemini 也能做到。或者，哦，我们一定是特别的。你也会这样想吗？这就是我听到你描述 AI 未来时的反应。

It is that we have to be at the center of the universe. Oh wait, we're not. We have to be the ones that are emotionally attuned. Oh wait, elephants have funerals. Oh we must be the ones that can be creative and create art. Oh wait, Gemini can do that. Or like, oh, we must be special. Do you find yourself doing that as well? That's my reaction as you're describing the future of AI.

是的，不。我认为我们是特别的，而且关于宇宙如何运作有很多深奥的谜团，包括我们头脑中的许多事物，也包括物理学中的事物。我想这就是为什么我很小的时候就决定从事 AI 研究，因为我在学校时就痴迷于那些大问题。通常，物理是我在学校最喜欢的科目，因为当你对所有大问题感兴趣时，你应该学习这门学科。但问题是，我在青少年时期阅读所有这些科学书籍和最优秀科学家的传记时——理查德·费曼是我一直以来的英雄之一——意识到尽管他们发现了很多，我们对世界了解很多，但还有太多我们不知道的东西。比如，我们不知道时间是什么。这对我来说太疯狂了。我们甚至无法描述如此基本的东西。我们沉浸其中，但它是什么？当然，它是熵之类的东西，但关于它到底是什么，并没有令人满意的解释。我们不太理解许多量子效应和引力，以及意识——实际上是我们关心的大部分事物。但大多数人整天用电视节目和游戏分散注意力，不太担心这些。我从来不是那样。这些深奥的谜团一直萦绕在我心头。我对最终关于现实本质的答案持相当开放的态度。我认为那才是我真正追求的，我想用 AI 作为工具来帮助我们理解现实的本质。无论答案是什么，我都相当乐观。从这个意义上说，我是一个真正的科学家：我对答案应该是什么没有任何预设的观念。我只想知道答案。

Yeah, no. I think we are special and I think there are a lot of deep mysteries about how the universe works, including a lot of things that are in our minds, but also things out there in physics. I think that's why I decided from a very young age to do AI, because I was obsessed when I was a kid at school with the big questions. Normally, physics was my favorite subject at school because that is the subject you're supposed to study when you're interested in all the big questions. But the thing was, I realized as a young teenager reading all these science books and biographies of the best scientists—Richard Feynman is one of my all-time heroes—that although they discovered a lot and we know a lot about the world, there's so much we don't know. Like, we don't know what time is. This is insane to me. We can't even describe something as fundamental as that. We're swimming in it, but what is it? Of course, it's entropy and things like that, but it's nothing satisfactory about what it really is. We don't understand a lot of quantum effects and gravity properly, and consciousness—actually most of the things we care about. But most people just distract themselves all day with TV shows and games and don't worry too much about it. I've never been like that. These deep mysteries play on my mind all the time. I'm quite open-minded about what the answers might be eventually about the nature of reality. I think that's ultimately what I'm after, and I want to use AI as a tool to help us understand the nature of reality. I'm quite sanguine about whatever the answer might be. I'm a true scientist in that sense: I don't have any prescribed notion of what the answer should be. I just want to know the answer.

我也是。

Me, too.

描述你正在做的事情的一种方式实际上是：创造一个系统，它不会特别擅长某件事或另一件事，而是像你一直说的那样，创造 AGI，即通用人工智能，它擅长所有事情。是的。我知道你是科幻迷。我也是。你能为我描绘一下你脑海中那部科幻电影的情节吗？那是一个你真正实现了这一目标的未来。

One way to describe what you're trying to do is effectively this: to create a system that wouldn't be especially good at one thing or another, but rather to create, as you've been saying, AGI, artificial general intelligence that would be good at it all. Yes. I know you're a fan of sci-fi. I am, too. Could you play out for me the plot of the sci-fi movie in your head that is the future where you actually do this?

是的，我可以。我也喜欢科幻，我小时候可能读得太多了——这也许能解释一些事情。我最喜欢的系列之一是伊恩·班克斯的《文明》系列。我认为它描绘了一个非常有趣的 AGI 后世界。他没有称之为 AGI，但那就是我们正在描述的，比如一千年后的未来。但我认为即使从现在起 50 年后，其中一些事情也可能发生，我们安全度过了 AGI 时刻。它被建造出来，对社会有帮助，它就在这里。也许我们甚至能把它放在口袋里。我们用它破解了一些我称之为科学中的根节点问题。AlphaFold 就是其中之一。这些问题，如果你把知识想象成一棵树，它们是根节点问题，一旦破解，就会解锁整个分支的新研究或新应用。我认为还有其他事情，比如核聚变，或者更好的常温常压超导体，然后你可以与最优电池结合。我认为能源问题会得到解决——以某种方式获得几乎免费的、可再生的清洁能源，核聚变或更好的太阳能。这将使我们真正能够星际旅行，因为太空旅行的主要成本仍然是火箭燃料，即能源成本。如果那几乎为零，因为我们破解了核聚变，可以从海水中制造无限火箭燃料，那么我们就可以到处建立催化剂工厂和海水淡化厂。这将解锁太空。然后我们将能够获得更多资源，因为我们可以开采小行星。所有这些属于科幻小说范畴的事情在未来 50 年内变得非常可能。围绕太阳的戴森球——水星恰好位于正确的位置，由正确的材料构成，这有点神奇。然后这有望带来人类的最大繁荣，帮助治愈所有可怕的疾病，让我们活得更长、更健康，并前往恒星，将意识带到银河系的其他地方。这将是一个惊人的结果，我认为在未来 50 年内可能发生。

Yeah, I can. I love sci-fi too, and I probably read too much of it when I was a kid—that might explain a few things. One of my favorite series was the Culture series by Iain Banks. I think it paints a really interesting post-AGI world. He didn't call it AGI, but that's what we're describing, like a thousand years in the future. But I think even 50 years from now, some of this could happen where we've gotten through the AGI moment safely. It's built, it's helpful for society, and it's here. Maybe we will have it in our pockets even. And we've used it to crack some of what I call root node problems in science. AlphaFold was one of those. These are problems where, if you think of the tree of all knowledge, they are root node problems which, if cracked, would unlock a whole branch of new research or new applications. I think there are other things like fusion, or better room temperature superconductors at atmospheric pressure that you could then combine with optimal batteries. I think there will be a solution to the energy problem—free, pretty much free renewable clean energy one way or another, fusion or better solar. And that will unlock us to really travel the stars because the main cost of space travel is still the rocket fuel, the energy cost. If that's sort of zero because we can make infinite rocket fuel out of seawater, having cracked fusion, then we can have catalyst plants and desalination everywhere. That unlocks space. Then we'll be able to get a lot more resources because we can mine asteroids. All these things that are the purview of science fiction become very plausible in the next 50 years. Dyson spheres around the sun—Mercury is conveniently in the right place and made of the right material, which is kind of amazing. And then that should hopefully lead to maximum human flourishing, help cure all these terrible diseases, so we live much longer healthier lives, and traveling to the stars bringing consciousness to the rest of the galaxy. That would be an amazing outcome, and I think could happen within the next 50 years.

我相信你。你说这些的时候，我喜欢你说的话，我相信你。至少这是我想做的。

I believe you. You're saying these things and I like when you're saying that my belief you. That's what I'm trying to do at least.

这是我的最后一个问题。如果我在自己的葬礼上像墙上的苍蝇一样，在人们说「她爱她的丈夫、家人和朋友」之后，我希望他们会说，她一生都在努力帮助人们看到乐观的未来，这样他们就能参与实现这些未来。他们可以更快地、更好地为更多人实现这些未来，或者无论人们决定用他们看到的愿景做什么。所以我的最后一个问题是：你希望他们怎么评价你？

This is my last question. If I were a fly on the wall at my own funeral after they said she loved her husband and her family and her friends, I would hope that they would say that she spent her life trying to help people see optimistic futures so that they can be part of making them happen. That they can make them happen more quickly or better for more people or whatever it is that people decide to do with the vision that they see. And so my last question for you is: what do you hope that they say about you?

我希望他们会说，我的一生对人类有益并服务于人类。这就是我正在努力做的。所以那将是最好的事情。

I would hope that they would say that my life was of benefit and service to humanity. That's what I'm trying to do. So that would be the best thing.

非常感谢你的时间。谢谢。真的很感激。

Thank you so much for your time. Thank you. Really appreciate it.

谢谢。太棒了。如果你想随时玩叠叠乐，我们可以玩。我们有一个修改版的叠叠乐。

Thanks. Awesome. If you want to play Jenga anytime, we can play. We have a modified version of Jenga.

你做得很好。所以，是的，这真的很棒。我不敢相信我们有这么多项目。当我看到那些积木时，真的很疯狂。所以，它们都立起来了。是的，它们上面都有我们的项目。你记住了所有东西的位置吗？好的。当然。所以，游戏就是把它抽出来，我们当时在玩这个。

You did that very well. So, yeah, this is actually awesome. I can't believe how many projects we've got. It's really crazy when I saw the bricks. So, they all got up. Yeah, they have all got our projects on them. Did you memorize where everything was? Okay. Of course. So, the game is you pull it out and we were playing this.

和你玩不公平，但你必须说出那个项目是什么。答错了不得分。

It's unfair to play with you, but you have to say what that project was. You don't get the point if you get it wrong.

比如，gnome，这是材料科学。

So, for example, it would be gnome, this is material science.

是啊，对你有点不公平。我本来希望我能赢，但你玩叠叠乐可能比我强多了。

Yeah, it's a little bit unfair on you. I mean, I would hope I would win this game, but you're probably a way better at Jenga than me.

好，我们试试这个。给你。AlphaCode，这个更清楚吧？Codeforces。

Yeah, let's see, let's do this one. There you go. Okay, AlphaCode. Yeah, that one's clearer, right? Code forces.

从某种意义上说，这是遗传学，但那是编码蛋白质的 2%？

From a sense, this is genetics, but the 2% that codes for proteins?

对。我们现在就得做。我有时间，可以把下一个推后。

Yes. We have to do this now. I've got time. I can push back my next...

太好了。等等，我还有一个问题。你知道 AlphaFold 吗？AlphaFold 是编码。对，可以用来编码。

Great. Wait, wait, I have one more question. You know AlphaFold? AlphaFold is coding. Yeah, we can be used for coding.

编程？它是将遗传算法与 Gemini 结合。这是我们在已知领域之外做类似 AlphaGo 事情的一次尝试。所以那个我不该得分。

Programming? It's combining genetic algorithms with Gemini. So, this is one attempt at doing like AlphaGo stuff beyond what is known. So, I wouldn't get a point for that one.

不，半分，半分。好吧，再问你一个问题。趁你还在，我就继续问，反正也没事。

No, half a point, half a point. Okay, one more question for you then. While I have you, I'm just going to keep going, because why not?

我们还在录。显然。

We're still rolling. Obviously.

好吧，有什么我没问但你认为人们应该知道的重要事情？

Okay, what did I not ask you that you think is important for people to know?

我没问什么？其实我们聊了很多。GenCast，这是天气预报。

What did I not ask me? I think we covered a lot actually. GenCast, this is weather prediction.

对。哦，我们没聊那个。纳维-斯托克斯方程，我完全忘了你们解决的那整个分支。

Yes. Oh, yeah, we didn't cover that. Navier-Stokes, I completely forgot about solving that whole branch of things.

我忘了你做的那个求解……所以，模拟是一个有趣的东西。我们没怎么聊那个，也没聊 Genie，模拟对 DQN 的作用，当然，一切从 Atari 开始。模拟可以帮助你理解某些科学领域，甚至社会科学如经济学，这些领域很难或无法进行实验，要么成本太高，要么无法进行对照实验。所以我一直很喜欢模拟。

I forgot about that thing you did solving... So, that was one interesting thing is simulations. We didn't talk much about that or Genie, which is the role of simulations to DQN, of course, started it all off, the Atari stuff. Simulations to help you understand some area of science or even social science like economics that you can't or are very hard to run, either expensive to run experiments or you can't run controlled experiments in. So, I've always loved simulation.

哦，对，我说了。我们俩都很好胜。所以这会很严肃。实际上，叠叠乐的规则是碰了就必须移动吗？

Oh, yeah, I said there we go. We're both very competitive, I think. So, this is going to be quite serious. Actually, in Jenga, are the rules that if you touch it, you have to move it, or not?

我们玩的是宽松版。

We are playing a loose version.

简单版。

The easier version.

而且因为我们在玩创意玩法，允许把它们推到一起，还可以用两只手。

Also, because we were doing a creative thing where you're allowed to push them together, you can use two hands also.

哦，好吧，通常不允许那样吧？我就拿这块。我又要用 AlphaCode 作弊了。

Oh, okay, you're not allowed to normally do that, right? I'm just going to take this one. I'm going to cheat with AlphaCode again.

我认为人们会问你的一个问题是，如果他们看了这个，对你描述的未来非常乐观，也有和你一样的担忧，他们听完后在想「我相信这个未来，我想参与其中」，你会如何建议他们参与？

One of the questions I think people will have for you is if they're watching this and they are very optimistic about the futures that you've described. They have all of your same concerns. They generally have gotten to the end of this conversation and they're thinking, 'I believe in this future and I want to be part of it.' How would you advise them to participate?

我在大学和学校演讲时会说，他们必须顺应潮流。我会让自己沉浸在所有可用的工具中，借助这些工具和能力变得几乎超级强大。因为我的印象是，即使在前沿实验室，我们也要投入大量工作来制作这些前沿模型的下一个版本以及所有相邻模型。所以，对我们来说，像 VO、Nano Banana 和 Gemini，我们只能探索你能用它做的应用的一小部分。而且我认为这个差距越来越大，能力过剩，最新模型上的所有酷东西，发布节奏也越来越快。所以，我认为对于那些真正擅长使用这些工具并将其应用到新领域的人来说，机会空间巨大。我觉得现在的孩子可能能用这些工具以某种没人想到的新方式创办一家数十亿美元的企业。像 Open Claw 这样的东西就是一个很好的例子。

I would, when I do talks at universities and schools, I would say they've got to just go with the flow of the direction. I would immerse myself in every tool available and just become almost superpowered with those tools and capabilities. Because my impression is even at the frontier labs, we have so much work that has to go into just making the next versions of these frontier models and all the adjacent models. So, for us like VO and Nano Banana and Gemini, we can only explore a fraction of the applied things you could do with it, the applications you could make with it. And I think that gap is getting bigger and bigger in terms of the overhang of the capabilities, all the cool stuff on the latest models, and the release schedules are getting faster and faster. So, I think the opportunity space is getting huge for people who are really expert at using those tools and then apply them to some new domain. I think a kid these days could probably start a multi-billion dollar business in some ways using these tools in some new way that no one had thought about. And I think things like Open Claw is a good example of that.

是啊。也许我们应该算平局，因为我觉得我们俩都输不起，对吧？所以，该你了。该你了。我们可以……我来试试我的内功。来吧。

Yeah. Maybe we should call it a draw, because I don't think either of us could bear to lose that, right? So, it's your move. It's your move. Yeah, it's your move. We can... I will try my inner move. Go on then.

你要让我出丑了。

You're going to make me make a fool of myself.

2016 年，你白板上有一张便利贴写着「解决蛋白质折叠」加笑脸。

In 2016, you had a sticky note on your board that said 'solve protein folding' smiley face.

对。

Yeah.

现在你白板上的便利贴写的是什么？

What is on the board now in your proverbial sticky notes?

回答你，我桌上有一堆大约一百张便利贴，所以……上面写什么？我其实不记得了。大概是一份今晚之前要完成的 30 件事的清单。

To answer it, I've got a pile of about a hundred sticky notes on my desk, so... What's on it? I can't actually remember. It will be a list of about 30 things that need to be done by like this evening.

我最好去处理它们。但看起来不错。我们是不是……你想……实际上，我会一直继续直到你停下。所以你可以随时叫停。

I better probably get to them. But look, great. Should we... Do you want to... Actually, I'm going to keep going until you stop. So, you can stop whenever.

好，我们……几点了？好吧，我再走一步。但现在我们有点作弊。我们在用已经……我最后要大胆一点……哦，加油。加油。加油。如果我拿到这块，我就再问一个问题。

Yeah, let us... What time is it? Okay, I'll do one more move. But now we're kind of cheating. We're using the pieces that already... I'm going to try and be ambitious in the last... Oh, come on. Come on. Come on. If I get this one, I get another question.

好，行。这似乎公平。哦，天哪。那怎么平衡？肯定不会。不。是的。

Yeah, okay. That seems fair. Oh, god. How is that going to balance? Surely not. No. Yes.

太棒了。谢谢。有这个主意真好。

That was awesome. Thanks. That was a great idea to have that.