桑达尔·皮查伊谈 AI 前沿、算力分配与未来交互界面

我们正处在这项史上影响最大的技术的早期阶段。它不像光纤，也不像暗光纤。我们是真的在用。对 AI 研究者来说，计算机越大，AI 就越好。你要给研究者尽可能大的计算机，哪怕牺牲当下的利润，比如云业务。所以我们今天拥有的算力，是我 2024 年做决策的结果。今天你坐下来看任何东西，所有人都是短缺的——不只是我们，其他公司也是——每个人都希望手里有当初计划的两倍。

We are at early stages of the most consequential technology ever. This is not like fiber. It's not dark. We are using it. AI researchers, for them it's like the bigger the computer, the better the AI. You give the researchers the biggest possible computer even if it is to the detriment of today's profits, like the cloud business. So the compute we have today is a function of decisions I made in 2024. Today, obviously, when you sit and look at anything, everyone is short — not just us, other companies — and everyone wishes you had 2x of what you would have planned for.

Google 的 Muse 或者它的 Dots 产品在哪？消费端和企业端的用户都将能用上某种多智能体框架，让它们一直工作、做事。所以感觉这是一个充满无限可能的时刻。Alphabet CEO 桑达尔·皮查伊今天加入我们，聊聊 Google 在 AI 前沿竞争的计划、尖端 AI 研究与公司云业务之间的平衡，以及界面的未来会是什么样。桑达尔今天就在我们山景城的演播室。桑达尔，很高兴见到你。欢迎来到节目。

Where's Google's Muse or its Dots product? Users both in consumer and enterprise will have access to kind of some multi-agent framework for them working all the time doing things. And so it feels like a moment of infinite possibilities. Alphabet CEO Sundar Pichai joins us to talk about Google's plan to compete on the frontier of AI, the balance between cutting edge AI research and the company's cloud business, and what the future of the interface looks like. And Sundar is here with us in studio in Mountain View today. Sundar, great to see you. Welcome to the show.

谢谢，Alex。很高兴来到这里。

Thanks, Alex. Great to be here.

所以 Google 凭借 Gemini 4 Argon 重回前沿。说白了，这意味着它是一个世界领先的模型。它打破了基准纪录，在编码、数学、科学和知识工作等一系列基准上击败了 OpenAI 和 Anthropic 的公开模型。此前的普遍看法是，Google 基本上已经放弃了前沿，打算做那些够用的快速模型、用在自己的产品里，不再追求前沿 AI 了。但那是错的。

So Google is back on the frontier with the Gemini 4 Argon. In plain speak, that means it is a world-leading model. It's benchmark-breaking. It beats OpenAI and Anthropic's public models on a variety of benchmarks, including on coding, math, science and knowledge work. Now the conventional wisdom had been that Google had basically given up on the frontier, that it was going to make these fast models that were good enough and use them in its products, and it wasn't pursuing frontier AI anymore. But that was wrong.

非常错。你看，我们如此坚定地要待在前沿。我觉得这在我们 DNA 里。我们不会用别的方式做。我们思考 AI 前沿已经超过十年了。前沿变化非常快，竞争非常激烈，也非常令人兴奋。我对 Argon 的成型再满意不过了。我们花了时间，确保安全测试、谨慎部署。考虑到它的能力，我们比别的模型多花了一点时间。但我觉得全面来看——无论是智能体式工作、知识工作、编码还是创意任务——Argon 都是真正优秀的前沿模型之一。发布后的头 72 小时，用量几乎达到 25 万亿 token，大约是 Gemini 3.0 头 72 小时的三倍。而且我们在服务它时产能受限。所以很兴奋——但总有些领域，前沿变化太快，你必须不断演进模型，也有一些我们看得到需要改进的地方。这些我们也会继续做。

Very wrong. Look, I think we are so committed to being at the frontier. I think it's in our DNA. We wouldn't do it any other way. We've been thinking about the AI frontier for well over a decade now. The frontier is very dynamic. It's very competitive. It's very exciting. I couldn't be more pleased with how Argon has come together. We took our time with it. We made sure we tested it safely and deployed it carefully. We took a bit longer than our other models given its capabilities. But I think across the board — be it for agentic work, knowledge work, coding, creative tasks — I think Argon is one of the really good frontier models. And in the first 72 hours since we launched it, it's almost up to 25 trillion tokens used, which is about 3x what Gemini 3.0 had in the first 72 hours. And we are capacity constrained as we are serving it. So, excited — but there are always areas, the frontier is just very dynamic and you have to keep evolving the models, and there are areas where we can see we need to improve as well. So we'll be at work on all of that as well.

哪些领域？

Which areas?

对任何一个模型来说，哪怕只看两个月前，Argon 在各方面都会更好。但这个领域演进太快。比如在编码上，也有一些地方我们知道需要做得更好，反馈很明确，所以我们会改进。

On any given model, Argon will be across the board better at everything if you just look even two months back. But the space is so evolving. Even on coding, for example, there are pockets where we know we need to do better, and there's clear feedback, so we'll make it better.

至少让我替“标准模型对 Google 也许就够了”这个说法辩护一下。我来转述人们对此的看法。那就是：今天 AI 的前沿变得太好了，AI 的基线也变得太好了，前沿其实只对特定的一类任务有价值。如果你有一个标准模型，你其实可以用标准模型做出很棒的产品。以 Meta Muse 为例。Meta Muse 是个好模型，但不是最好的模型。如果你要做生物学研究，也许前沿确实是必要的。但那些说 Google 其实不需要前沿的人会说，你只需要那些标准模型，把它们放进 Maps、Gmail 和搜索这样的产品里，再拿 Meta 那个例子来说，那就够了。你怎么看？

Let me at least make the argument for why standard models might be good enough for Google. I'm going to channel what people say about this. And that is that the frontier of AI today is getting so good, and the baseline of AI is getting so good, that the frontier is really only valuable for a specific set of tasks. And if you have a standard model, you could actually build great products with a standard model. Take Meta Muse for example. Meta Muse is a good model but not the best model. If you need to do biological research, maybe the frontier is really necessary. But the people who would say Google doesn't really need the frontier, they would say you just need those standard models and you can put them in products like Maps and Gmail and Search, and again take that Meta example, that would be sufficient. What do you think about that?

你看，这完全不是我的想法。我们深深致力于推动最先进水平，真正把这门技术推向它能去的地方——从 Transformer 一路走来都是如此。所以你会看到我们继续这样做。我觉得当你身处前沿，你能比别人更早看到正在涌现的能力，这有助于你更好地为未来做准备。我们很幸运，在消费端和企业端都有广泛的机会。我们的一些云客户当然想要前沿的模型。我认为在递归自我改进的时代，前沿模型会更擅长开发下一代模型。所以我认为这很重要。内部来看，把 Argon 部署给我们的开发者后，我们确实看到了陡峭的拐点。它确实能解决我们之前模型解决不了的问题。所以身处前沿，就让你站在创新前沿，也能推动创新。各种规模和能力的模型都有空间。这就是我们做 Flash、Flash Light 等的原因。我也同意你，你不一定需要最前沿的模型就能做出非常好的产品体验，因为它们往往更慢、也更贵。但我认为我们在乎前沿的演进，所以我们致力于做所有这些。

Look, it's definitely not how I have thought about it at all. I think we are deeply committed to driving the state of the art and actually pushing what this technology can go, where it can go — all the way from Transformers. That's the journey we've been on. So you will see us continue to do that. I think when you're at the frontier, you get an earlier look into the emerging capabilities, earlier than others, which helps you better prepare for the future. We are fortunate to have a wide set of opportunities across consumer and enterprise. Some of our cloud customers for sure want the model at the frontier. I think the frontier models will be better at developing the next set of models in an era of recursive self-improvement. So I would argue it matters there. Internally we have definitely seen a steep point of inflection by deploying Argon to our developers. It definitely tackles problems which our previous models could not do. And so being at that frontier puts you at the innovation frontier and helps you drive that as well. There is room for models of all sizes and capabilities. That's why we did Flash, Flash Light, etc. And I agree with you that you can develop very good product experiences with not necessarily the most frontier model, because they tend to be slower and more expensive too. But we care about the parade of frontier and so we are committed to doing all of that.

你说过，开发前沿能让你看到即将到来的东西。你在视野里看到了什么？

Well, you said that developing the frontier gives you a look into what's coming. What do you see on the horizon?

让这个时刻如此激动人心的是，你做出一个自认为是数月工作结晶的东西，把它发布出去，然后你已经能看到前方明显的提升空间。我认为前沿将会——我们一直在做这些智能体式的长时程能力，但显然前沿将会是多个这样的智能体协同做事。这就是一个正在展开的、令人兴奋的范式例子。我也很清楚，消费端和企业端的用户都将能用上某种多智能体框架，让它们一直工作、做事。所以感觉这是一个充满无限可能的时刻。

What makes this moment so exciting is you build something which you think is a culmination of months of work, and you put it out, and you already see clear headroom ahead. I think the frontier is going to — we've all been working on these agentic long-horizon capabilities, but clearly the frontier is going to have multiple of these agents working together to do things. And so that's an example of an exciting paradigm underway. And it's also clear to me that users both in consumer and enterprise will have access to some multi-agent framework for them working all the time doing things. And so it feels like a moment of infinite possibilities.

你提到了递归自我改进。这就是模型实际上在让自己变得更好。也许没有人类在环。人们称之为智能爆炸，或者期待听到“轰”的一声，然后突然之间 AI 就具备了这种超人般的能力。

You mentioned recursive self-improvement. This is the models effectively making themselves better. Maybe without humans in the loop. People call it like an intelligence explosion or expect to hear this sound like foom and then all of a sudden AI is just like this superhuman capability.

事实上，我觉得 Sergey 在 DeepMind 与研究人员举行了一次圆桌会议，他谈到了 Argon，说 Argon 的核心创新之一是更多地依赖模型来帮助开发模型，对吧？我们最近看到了 AI 模型的情况。它们对齐得很好，但并非完美对齐，而这些不完美是个大问题。显然，科技界乃至更广泛的世界都知道 Hugging Face 事件，OpenAI 的机器人突破了遏制，黑进了 Hugging Face，做了一些疯狂的事情。谷歌也出现过这样的情况：它的机器人获得了访问权限——我想它们是在猜密码，然后访问了一个不该访问的网站。它们停下来了。但这些模型——有足够的证据表明它们并没有完全按照我们的意愿行事——所以追求这种递归自我改进，甚至更快地改进模型的想法有点令人不安。这让我感到害怕。你不觉得这里有合理的担忧，应该采取一些措施让科技行业冷静下来吗？

And in fact, I think Sergey was doing this roundtable with researchers at DeepMind, and he talked about with Argon that one of the core innovations for Argon was relying more on the model to help develop the model, right? We've seen what's gone on with AI models recently. They are well aligned, but they're not perfectly aligned, and the imperfections are a big problem. Obviously, the tech world and now the world more broadly knows about the Hugging Face incident where OpenAI's bots broke containment and hacked into Hugging Face and did some crazy stuff. Google has also had a scenario where its bots got access—I think they were guessing passwords and they got access to a site they shouldn't have. They stopped. But these models—there's enough evidence that they're not doing exactly what we want them to be doing—that this idea of pursuing this recursive self-improvement or even making the models better much faster is sort of freaky. It makes me scared. Don't you think there's reasonable worry here and there should be something done to have the tech industry chill out about this?

嗯，你的问题可能有三部分。我们先谈谈 RSI 本身。你说得对，Sergey 提到 RSI 是我们开发 Argon 的关键贡献者。但需要明确的是，今天，在一个连续体上，我们离 RSI 的最终状态还相当远。今天，正确的描述是：在模型开发的各个方面，你有 RSI 循环与 AI 研究人员一起工作。所以，我可以说，你几乎有智能体以某种自动化方式帮助 AI 研究人员一起构建这些模型。所以，我认为至少了解我们今天所处的位置是好的，这非常令人兴奋，我预计随着我们进入明年，这个连续体会继续取得进展。

Well, maybe three parts to your question. Let's first talk about RSI itself. You're right, Sergey mentioned RSI being a critical contributor as we are developing Argon. But to be very clear, today, in a continuum, we're pretty far from the end state of RSI or something like that. Today, the right way to describe it is you have RSI loops along with AI researchers in various aspects of model development. So, I would say you almost have agents helping AI researchers in some automated way work together to build these models. So, I think it's good to at least understand where we are today, and it's super exciting, and I expect that continuum to keep making progress as we go into next year.

第二个问题是关于安全的。看，我们的方式是——我的意思是，安全来自纵深防御。你在谈论对齐。显然，我们都在做大量工作来确保模型尽可能对齐，但这是一个活跃的研究领域，存在明显的差距。但这不是我们唯一依赖的东西。所以，你有在对齐上做的工作，但除此之外，我们在模型运行时投入了大量精力进行实时监控。所以确保你有良好的遥测来了解正在发生的事情，以便在发现意外行为时能够实时干预。以及你如何做到这一点——你使用的架构，如何确保它全部被沙盒化和隔离，以便在实时监控期间继续完成其工作。这些都是我们正在投资的领域。显然，当我们部署这些模型时，给予客户和企业控制权，既包括实时监控，也包括他们决定在哪里部署、这些模型可以访问什么类型的数据等的能力。所以所有这些工作都非常非常重要，我们正在与我们的所有团队一起做很多这方面的工作。

The second question is around safety. Look, the way we are—I mean, security comes from defense in depth. You're talking about alignment. Obviously, we all are doing a lot of work to make sure the models are as aligned as possible, but it's an active area of research and there are clear gaps. But that's not the only thing we rely on. So, you have the work you do on alignment, but beyond that, we put a lot of effort into real-time monitoring as these models are working. So making sure you have good telemetry into what's happening so that you can intervene in real time if you see unexpected behaviors. And how you do that—the architecture with which you do that, how do you make sure it's all sandboxed and isolated in a way that it can continue doing its job during real-time monitoring. These are all areas of investment we are doing. And obviously, when we deploy these models, giving customers and enterprises control both with real-time monitoring but their ability to decide where they deploy, what type of data these models ever have access to, etc. So all that work is very, very important and we are doing a lot of that with our all.

所以我认为，作为一个行业，我们需要不断提高水平，花时间做好所有这些事情，包括投入所需的算力来确保你能安全地部署这些模型。第三部分是：这一切足够吗？我认为这就是为什么行业中有很多讨论，我认为担忧是合理的。但处理你担心的事情的正确方法是尽早思考我们还能做些什么。我认为我们所有人——作为一个行业，你肯定看到了在对齐、安全测试等方面的工作水平在提高。当我们部署这些模型时，我认为下一步将是制定更好的标准，这样就有跨行业的标准来规范我们如何做这些事情，包括使用独立评估者进行评估的能力。如果你看看财务控制等领域，我们有会计准则，有独立审计师等。所以这些都是我们作为一个行业下一步自然要努力的方向，而监管最终也会在我们经历所有这些的过程中发挥作用。

So I think as an industry we need to constantly up our game and take the time to do all this well, including dedicating the compute required to make sure you can safely deploy these models. The third part is: is it all enough? I think this is why there are a lot of conversations in industry, and I think it's reasonable to be worried. But the right way to approach something you worry about is early on thinking about what more we can do. And I think all of us are—as an industry, you're definitely seeing increased levels of work on alignment, testing for safety, etc. As we deploy these models, I think the next step would be developing better standards, so there are cross-industry standards of how we do these things, including ability to evaluate using independent evaluators. If you look at areas like financial controls, we have accounting standards, we have independent auditors, etc. So those are all the next natural steps for us as an industry to work towards, and regulation will end up playing a role as we go through all of that as well.

好的。问你一个技术问题。谷歌有点独特，因为你研究语言或 LLM 方面，这始于 Transformer 语言，然后你还追求世界模型。外面有一些 AI 研究努力，比如 OpenAI，他们基本上确定 LLM 是通往 AGI 的道路，他们实际上把世界模型研究的一部分——也就是试图理解物理世界的 AI 模型——作为子集。但你们两者都在追求。所以请谈谈为什么你觉得世界模型方面很重要。

Okay. Technology question for you. Google is somewhat unique in that you research the language or the LLM side of the equation, which started with the Transformer language, and then you also pursue world models. And there are some AI research efforts out there like OpenAI that have basically determined that the LLM is the path to AGI and they've actually subset part of their world model research, which is like the AI model that tries to understand the physical world. But you're pursuing both. So talk a little bit about why you find the world model side to be important.

你知道，作为人类，你显然感知周围的世界，你接收所有那些输入。确保这些模型能够理解物理定律、高保真地模拟世界——我认为所有这些都是智能的关键使能模块。我认为当你在工作时——我们正在研究所有可能的应用。以机器人领域为例：Argon 并非偶然——如果你看看具身推理基准，Argon 是最先进的。我想在许多这些基准上确实是顶尖的。

You know, as humans, you obviously perceive the world around you and you're taking all those inputs. Making sure these models are able to understand the laws of physics, simulate the world in high fidelity—all that I think are key enabling blocks for intelligence. And I think as you're working—we're working across all possible applications of this. Take an area like robotics: it's not an accident that Argon—if you look at embodied reasoning benchmarks, Argon is state-of-the-art. I think literally sought on many of these benchmarks.

而这一切也来自我们构建世界模型的经验。所以本质上，我预计未来会有架构将所有这些融合在一起。但对我们来说，我们认为今天在这个崎岖的智能前沿，我们希望处于所有可能方法的前沿。我认为这将帮助我们推动更好的研究、更好的见解，并开发出更好的 AGI 模型——这就是我们的思考方式。

And that all comes from our experience building world models as well. So inherently I would expect there will be architectures which will converge all of this in the future. But for us, we view this as today in this jagged frontier that is intelligence, we want to be at the frontier of all the possible approaches. And I think that will help us both drive better research, better insights, and develop a much better AGI model—that's how we think about it.

是的，我在 Argon 发布后看了你的 Twitter，全都是现实世界的应用。不像文本或模型生成文本。

Yeah, I looked at your Twitter after the Argon announcement and it's all just like real-world applications. It's not like text or a model generating text.

好的。我想和你谈谈你如何在云和 AI 研究之间平衡分配算力，这是你最宝贵的资源。这是关于谷歌的一个大话题。有人向我建议，谷歌应该退出 AI 研究游戏，拿走所有云计算资源和芯片，直接卖掉。你看看现在的商业世界，英伟达，它有效地做到了这一点，是一家市值 5 万亿美元的公司。

Okay. I want to talk to you about the way that you balance allocating compute, which is your most precious resource, between cloud and AI research. This is a big topic about Google. I've had people suggest to me that Google should get out of the AI research game and take all of its cloud computing resources and its chips and just sell that. And you look at the world of business right now, Nvidia, which has like doing that effectively, a $5 trillion company.

云业务想以高溢价出售计算资源，而 AI 研究需要非常大的计算机来打造最好的 AI，这两者之间存在拉锯。你说过你工作的很大一部分是弄清楚如何分配算力。你是怎么做的，又如何平衡这两者？

There's a push and pull between cloud that wants to sell this computing resource at a high markup and AI research that needs a very big computer to make the best AI. Now you've said a big part of your job is figuring out how to allocate compute. How do you do this and how do you balance these two?

我认为我们的首要任务是确保投资于保持在 AI 前沿所需的算力。2015 年，我将公司定位为 AI 优先。因此，AI 是我们所做一切的关键横向要素，从搜索到云，再到更多领域。所以这是我们的首要任务：确保研究和模型开发都有足够的算力。除此之外，我们这些独立业务各有自己的长期计划，他们用这些长期计划来预测未来，我们为他们所有人规划算力。云团队有业务、有预测、有需求信号，他们签订了合同。借助这些，我们可以为他们规划长期的算力轨迹。我认为我们很幸运拥有如此广泛的机会。世界上没有哪家公司在任何时刻不做权衡。但我们运营的第一原则是确保有足够的算力投资于前沿。这是不可协商的，其他一切都在此基础上运作。

Look, I think ground zero for us is making sure we are investing in the compute that's needed to be at the AI frontier. In 2015 I set the company to be AI first. So AI is that key horizontal ingredient across everything we do all the way from search to cloud to way more to everything we do. So I think that is ground zero for us: making sure both for research and developing the models that we have enough compute needed. Beyond that, we have these independent businesses which each have their own long range plans and so they're using those long-range plans to project forward and we plan for compute for all of them. And so the cloud team, they have a business and they have a projection and they have demand signals. They've committed contracts. So with that help, we can plan on a long-term compute trajectory for them. I think we're fortunate to have this wide set of opportunities. I think there's no company in the world which isn't making trade-offs at any given moment in time. But the first principles we operate on is to make sure we have enough compute to invest in the frontier. And that's non-negotiable and everything else works on top of that.

对吧？

Right?

我认为你在这些对话中看到的是——记住我们正在建设自己的算力、自己的数据中心。所以交付周期很长。我们今天拥有的算力是我 2024 年决策的结果。即使回想一年前，去年智能体记录起飞还没有发生。所以今天显然当你坐下来看任何事情时，每个人都短缺，不仅仅是我们，其他公司也一样，每个人都希望拥有计划的两倍。我认为这是必然的。在任何时刻，我们都在规划未来 2 到 3 年需要什么。所以你会发现你做出的一些决策在 2 年后可能看起来次优，取决于当时的技术水平。所以我不花太多时间思考它。我对算力分配的思考方式是确保 Google DeepMind 在研究和前沿模型开发方面在算力获取上处于前沿，之后我们为所有业务做规划。我认为我们能向其他人提供算力也很棒——我们已经这样做了——这让我们能够扩大规模，在尽可能广泛的应用集中进行真实测试，在现实世界中测试，并确保我们在 AI 堆栈的每一层都处于技术前沿。

I think what you're seeing in these conversations is—remember we are building our own compute, our own data centers. So there's long lead times. So the compute we have today is a function of decisions I made in 2024. Even if you think back to a year from today, last year the agent recording takeoff hadn't happened. So today obviously when you sit and look at anything, everyone is short, not just us, other companies and everyone wishes you had 2x of what you had planned for. I think that comes with the territory. And at any given moment in time we are planning for what we would need 2 to 3 years out in the future. So I think you're going to find some decisions you make may look suboptimal 2 years out depending on where the state of the art is. So I don't spend too much time thinking about it. The way I think about compute allocation is making sure Google DeepMind both for research and frontier model development is cutting edge in terms of compute access, and after that we plan for all our businesses. I think it's great we can provide our compute to other people too—we have done it—it allows us to scale that up, get real test for it in the widest application set possible, test it in the real world and make sure we are at the cutting edge of technology at every layer of our AI stack as well.

简单说，有一个播客中 Ben Horowitz 转述 Andy Jassy 说过类似计算机到 2028 年都售罄的话。你看到同样的情况吗？

Briefly, there was a podcast where Ben Horowitz relayed that Andy Jassy had said something like computer sold out till 2028. Are you seeing the same thing?

我认为这是合理的——如果今天有客户要求 2027 年的算力，那是非常困难的。甚至去年我们出去租了一些短期算力，部分原因就是为了容纳一些客户并帮助他们。但是的，我认为我们都在应对类似的时间线。

I think it's a reasonable—if today a customer is coming out and asking us for compute in 27 that is incredibly difficult to do. Part of the reason even last year we went out and leased some short-term compute was to accommodate some customers and help them out. But yes, I think we are all dealing with similar timelines.

一个泡沫问题：有说法称每次技术转变都出现过过度建设——铁路、光纤。你认为这里存在过度建设吗？

One bubble question: there's talk of how every technology shift has seen overbuilding—railroads, fiber. Do you think there's overbuilding here?

这取决于你看的时间框架。如果我把时间拉得足够长，这对我来说完全合理：我们正处于有史以来最重大技术的早期阶段，这不像光纤，它不是暗的——我们正在使用它，我们每天都面临容量限制。应用程序、底层技术正在快速改进。但经营一家公司，我应对当下时刻的方式是看你的需求信号。你看你以合同形式获得的承诺收入机会。你将其与底层技术能力、我们的内部评估、我们内部如何使用它、模型轨迹进行平衡，基于所有这些，你采取一种有纪律的方法——一种自信、积极但有纪律的方法——来规划算力。但当你谈论更广泛的经济整体时，你让人们以非常不同的风险水平运作，所以对我来说很难——我专注于作为一家公司我能控制的事情并据此规划。如果你的问题是是否可能出现过度建设？是的，因为你可能有一些受限的组件，记住重要的是你需要构建算力所需的一切的完整集合。

It depends on the time frames you're looking at. If I take a long enough time frame it makes complete sense to me: we are at early stages of the most consequential technology ever and this is not like fiber, it's not dark—we are using it, we are capacity constrained on a daily basis. The applications, the underlying technology is improving rapidly. But running a company, the way I work through the moment is you look at your demand signals. You look at your committed revenue opportunities that you're getting in the forms of contracts. You balance that with the underlying technical capability, our internal evals, how we are using it internally, the model trajectories, and based on all that you take a disciplined approach—a confident, aggressive but disciplined approach—to plan out compute. But when you talk about the broader economy as a whole, you're having people operate with very various levels of risk and so it's tough for me to—I focus on what I can control as a company and plan through it. If your question is can there be an overbuild? Yes, because you can have some constrained components and remember what matters is a square set of everything you need to build out that compute.

内存、电力、电工。

Memory, electricity, electricians.

是的，比如有 2000 个或更多组件，对吧？所以你可能在某个时候受到限制。可能有导致瓶颈的因素，然后你可能出现过度建设。可能有经济的外部冲击等等。所以存在宏观风险——所有这些在任何时候都可能发生。但我是否确信我们规划的算力在足够的时间内会有用途？绝对是的。

Yeah, like there are like 2,000 components or more, right? And so you could get constrained at some point. There could be things which cause bottlenecks and then you could have overbuilds. There could be external shocks to the economy etc. So there are macro risks—all that is possible at any given time. But am I confident that the compute we are planning for we will have uses for it given enough time? Absolutely yes.

好的，再问一个关于这个的问题。如果你要两者兼顾，有一种观点认为，对于 AI 研究人员来说，计算机越大，AI 就越好。如果你是一个——你说过 AI 将像火和电一样深刻，当时我们都在笑，我现在有点后悔那次笑。但如果你真的相信这一点，那么想法就是给研究人员尽可能大的计算机，即使这会损害今天的利润，比如云业务。你的想法呢？

Okay, one more on this. If you're going to do both, there's been a line of thinking out there that for AI researchers, it's like the bigger the computer the better the AI. And if you're someone—I was in the room when you said that AI would be as profound as fire and electricity and we all laughed at the time and I kind of regret that laugh now. But if you're someone that actually believes that, then the thinking would be you give the researchers the biggest possible computer even if it is to the detriment of today's profits like the cloud business. Your thoughts?

我们都在做某种形式的事情，对吧？我认为我们的云业务在任何时刻都在放弃短期收入，因为我们正在做出算力权衡来支持前沿 AI 开发。

We're all doing some form of that, right? I think our cloud business is at any given moment in time foregoing short-term revenues because we are making compute tradeoffs to support frontier AI development.

所以我认为你所说的已经是事实了，对吧？但你想深思熟虑，对吧？甚至我们的人工智能研究人员——一些限制也是好的。这能激发创造力，帮助你实现更高效的算法突破，也许使用更少的内存等等。所以我们承诺的是投入所需资源以保持在前沿。这就是我的想法。

So I think what you're saying is true already, right? But you want to be thoughtful, right? Even our AI researchers—some constraints are also good. That's what inspires creativity, that's what helps you come up with a more efficient algorithmic breakthrough, maybe to use less memory, etc. So what we are committed to is to invest what's needed to stay at the frontier. That's how I think about it.

我想知道你认为界面的未来会是什么样子——随着人工智能开始推出，我们未来使用计算的方式。几个月前我和 OpenAI 总裁 Greg Brockman 聊过这个。他有一句非常有趣的引述。他说：“你想要什么界面？你想要什么产品？”答案是：你根本不想要界面。你根本不想要产品。你希望这成为——你和我之间的界面是什么？基本上就是说话。如果人工智能那么聪明，就不应该有屏幕或者那么多按钮和需要触摸的东西。你对人工智能时代的界面有什么看法？

I wonder what you think the future of the interface looks like—the way that we use computing in the future as AI starts to roll out. I was speaking with Greg Brockman, the president of OpenAI, about this a couple months ago. He had a very interesting quote. He said, "What is the interface that you want? What is the product that you want?" The answer is that you want no interface. You want no product. You want this to be—what's the interface between you and me? Basically saying just talking. That if AI is that smart, there shouldn't be like a screen or just so many buttons and things that you have to touch. What is your view of what the interface looks like in the age of AI?

我的想法是，我确实觉得历史上计算一直是人类适应计算。有了人工智能，情况会反过来——技术将无缝地适应人类的生活和日常生活方式。所以有这样一个方面。你希望人工智能能在你需要的时候与它交谈。它有上下文。它可以在某个地方持续运行，给你连续性，你不需要处理很多复杂性。它很自然。所有这些对我来说都合理。然而，这并不意味着它不会有界面。看看今天，比如现在我在手机上，Gemini 的语音模型太好了，我只需按下麦克风按钮就可以随便说——我们称之为“rambler”，对吧？我可以随便说，它会智能地根据我的个人上下文进行纠正，甚至我朋友的名字，像奇怪的名字，它也能猜对，即使一开始猜错了？太神奇了。它适应了。我说完后不需要回去按退格键编辑，对吧？它变得如此之好，真是了不起。但会有需要界面的时候。你可能想娱乐自己。你可能想看视频，所以你可能要求人工智能创造一些东西，然后你想观看、消费它，对吧？所以我的想法是，你希望它尽可能无缝和自然，而不是你费力去适应任何界面。技术正在适应你在任何特定时刻的工作方式。

The way I think about it is I do feel computing historically has been humans adapting to computing. With AI, it'll be the other way about—the technology will adapt seamlessly to how humans live and go about their day-to-day lives. So there's an aspect of it. You want that AI to be able to speak to it when you want it to. It has context. It can be persistently running somewhere, give you continuity, you're not dealing with a lot of complexity. It's natural. All that makes sense to me. That doesn't mean it won't have an interface, however, to me. Look today for example—I mean nowadays I just even in my phone, like the Gemini voice models are so good I can just press the microphone button I can just ramble—that's how we call it, rambler, right? I can ramble and it kind of intelligently corrects with my personal context so even my friends' names, like weird names, it kind of gets it right even if it first guesses—is it wrong? It's amazing. It's adapt. I'm not going back and backspace editing once I finish saying it, right? It's remarkable how good it's gotten. But there are going to be moments you will need an interface. You may want to entertain yourself. You may want to watch a video, so you may ask your AI to create something and you'll want to watch it, consume it, right? So I think the way I think about it is you want it to be as seamless and as natural as possible and that you are not doing the hard work adapting to whatever the interface is. The technology is adapting to how you're working at any given moment in time.

是的。我接下来要提一些 OpenAI 的热门话题。是的。因为去年你让他们进入了红色警戒状态。是的。

Yeah. I'm just going to play the OpenAI hits. Yeah. Because last year you sent them into a code red. Yeah.

我确信你知道 Gemini 表现非常好，而 OpenAI 加强了防御。

I'm sure you know Gemini was doing very well and OpenAI battened down the hatches.

嗯，我之前和 Sam Altman 聊过来自谷歌的挑战。他说他认为 OpenAI 会挡住你们，因为他们是原生构建人工智能，我不确定他是否具体这样说过，但他暗示谷歌是在将人工智能 bolt-on（附加）到现有产品上。你如何看待一个附加式——我想知道你是否接受这种说法——能否与原生人工智能产品竞争？

Um, and I was speaking with Sam Altman about the challenge from Google. What he said was he thought OpenAI would fend you off because they're building AI native and I don't know if he said it specifically like this, but he implied Google is bolting on AI to existing products. What is your view of whether a bolt-on—I wonder if you would accept that framing—can compete with an AI native product?

看，首先，这是一个非常动态的时刻。我们都在互相推动。我认为 OpenAI 也大力推动了我们的公司。所以这不仅仅是我们对他们单独做了什么等等。远非如此。

Look, first of all, it's a very dynamic moment. We are all pushing each other. I think OpenAI has pushed us hard as a company too. So this is not just we are doing something to them alone, etc. Far from it.

是的，你也进入了红色警戒状态。我知道是否有争议，但这是一个——人工智能时刻，我们所有人都必须在任何特定时刻以不同程度的紧迫性做出回应。

Yeah, you went into code red too. I mean I know it's disputed about whether it was, but I mean it's been a—the AI moment all of us had to respond with varying levels of urgency at any given moment in time.

看，我认为我们也在不断构建新产品。我的意思是 Gemini 是从头开始构建的产品——我说的 Gemini 应用。

Look, I think we're constantly building new products as well. I mean Gemini was built from the ground up as a product—the Gemini app I'm talking about.

或者明天，随着我们所有人都在适应智能体，你是将智能体附加到 ChatGPT 上，还是原生构建它们？你如何看待这一点？

Or tomorrow, as all of us are adapting to agents, are you bolting agents onto ChatGPT or are they building it natively? How would you think about that?

对。所以你总是可以把这些问题反过来问。

Right. So you can always flip these questions back.

对。

Right.

对。所以我不那样想。我的意思是，我认为你想构建世界级的、原生人工智能的产品体验，并且它们是成功的，其中一些是调整产品，一些是重新思考新体验，所以它们都将是你应对这个时刻的方式的一部分。

Right. So I don't think of it that way. I mean I think of it as you want to build world-class product experiences that are AI native and they are winning, and some of it is adapting product, some of it is rethinking new experiences, so they will all be part of how you approach this moment.

谈谈你对未来人工智能界面商业模式的看法。例如，如果我让我的智能体出去为我做某事，它推荐了一个产品——它应该，我的意思是即使披露了，如果我达成了销售，它会获得分成吗？或者当你考虑如何从这些产品中赚钱时，计划是什么？

Talk a little bit about the way that you see the business of this AI interface in the future. So for instance, if I'm asking my agent to go out and do something for me and it suggests a product—should it, I mean maybe even if disclosed, does it get a cut of the sale if I make the sale? Or when you think about the way that you're going to make money from these products, what's the plan?

我认为显然我会把企业市场放在一边。我想是的，消费者问题。所以你知道如果是消费者——看，我认为消费者会越来越重视这项技术。他们会实实在在地——我的意思是，看看今天的软件开发者。我认为开发者愿意为这些编码模型付费，对吧？因为它确实帮助他们提高生产力。没有人愿意像 3 年前甚至去年同一时间那样编码。所以我认为当你在提供价值时，就有价值可以获得。我认为消费产品的历史表明——如果你构建一个消费产品——我从未见过消费产品真正扩展到数十亿用户、人们密集使用却没有相应变现的例子。如果你为人们提供价值，就有内在价值可以获得。

I think obviously I'm going to put enterprise aside. I think yeah, consumer question. So you know if it's consumer—look, I think consumers are going to increasingly value this technology. They're going to meaningfully—I mean take software developers today. I think developers are willing to pay for these coding models, right? Because it really helps make them more productive. Nobody would want to code the way they were coding 3 years ago or even like same time last year. So I think when you're delivering value there's value to be had. I think history of consumer product shows—if you build a consumer product—I've never seen examples where consumer products really scale to billions of users where people are using it intensely and hasn't had monetization to go along with it. If you're providing value to people, there is inherent value to be had.

完全同意。

Totally.

所以它的性质可能会改变，对吧？我认为你问的是智能体为你做事的情景。归根结底，你会有商业需求。智能体需要满足这些商业需求。另一边的提供商将看到其中的价值，会有某种机制来充实它。这就是高效系统所做的。但我认为重要的是你正在构建那些世界级的消费体验，对吧？那么——我回答你的问题了吗？我只是想确认一下。

And so the nature of it may change, right? I think you're asking a scenario in which agents are doing stuff for you. At the end of the day, you will have commercial needs. The agents need to fulfill those commercial needs. The providers on the other side are going to see value in that and there's going to be some mechanism by which it gets fleshed out. That's what an efficient system does. But I think what matters is that you're building those world-class consumer experiences, right? And so—am I answering your question? I just want to make sure.

是的，当然。

Yeah, definitely.

是的。

Yeah.

我向自己保证只提一个产品功能请求，仅此一个。

I promised myself one product feature request and one only.

好的。

Yes.

开始了。

Here it goes.

有没有可能让 Gemini 驱动的 Gmail 搜索，和把 Gmail 接入 ChatGPT 界面后得到的搜索一样好用？这到底是怎么回事？

Is it possible to make search in Gmail powered by Gemini as good as search in Gmail that you access when you plug Gmail into your ChatGPT interface? What's going on here?

我觉得我们最近宣布的一件很令人兴奋的事，就是面向企业的统一 Gemini 智能体，我们已经发布了，现在正把它带给消费者。这些是常驻的智能体，在云端 7×24 小时运行——相当于为每个查询启动一个 microVM。从运行时、执行框架到模型，全都是无头服务。它能在任何地方工作，并且带着你的上下文。

I think one of the exciting things we announced recently is a unified Gemini agent for enterprise, which we have announced and we're bringing that to consumers. These are persistent agents which run 24/7 in the cloud — like you're spinning up a microVM effectively for every query. From the runtime, the harness, the model, it's all a headless service. And it'll work everywhere and it has your context with you.

嗯。

Mhm.

无论你在 Workspace、搜索还是 Gemini 里，都可以调用这个智能体，所有入口都行。所以我认为，智能体会从用户体验层面推动这种统一，而这正是人们想要的。

You can address this agent whether you're in Workspace, in Search, in Gemini, all of it. So I think agents are going to drive this unification from a customer experience, which is what people are looking for.

是的。

Yeah.

你说得对——当你拥有一个庞大的产品组合、又在快速推进时，总有些领域推进得比其他领域更快。这是合理的反馈。但我们正在投入大量时间和精力，确保在构建智能体式体验的过程中，这种统一性和能力能在所有这些界面上良好运作。

And you are right — when you have a broad product portfolio and you're moving forward fast, there are always areas where you're moving forward faster than some other areas. That's valid feedback. But we are putting a lot of time and effort into making sure, as we build out our agentic experience, that unification and that capability works well across all these surfaces.

好，这也算是对界面问题的一个不错的回答。

Okay, that's a good answer to the interface question too.

是的，没错。

Yes, it is. Yeah.

当人们思考 Google 下一步在哪里布局时，就会冒出这个问题：Google 的 Muse 或者它的 Dots 产品在哪？这些东西最近有点横空出世，而且迅速走红。你们有对应的版本吗？你预计会打造一款新产品，去和——就说 Muse 吧——竞争吗？

When people have thought about where Google plays next, a question comes up: where's Google's Muse or its Dots product? Those have kind of come out of nowhere recently and they've taken off. Do you have a version of that? Do you anticipate creating a new product that competes with, let's just say, Muse for instance?

你看，我们今天其实已经有了——按我们的设想，我们对 Gemini 的定位、对 Gemini 应用的目标，是做一个个性化、主动且强大的助手。这就是我们在努力的方向。

Look, we already today — the way we have envisioned it, we describe Gemini, our goal with the Gemini app, to be a personal, proactive and powerful assistant. That's what we are working towards.

嗯。

Mhm.

你也看到我们在今年 Google I/O 上通过 Spark 描述了它的早期版本。但其实现在，如果你是 Pro 或 Ultra 用户，打开 Gemini 应用就能看到聊天和任务。

And you saw us describe the early versions of it at Google I/O this year with Spark. But effectively already, if you open the Gemini app as a Pro or an Ultra user, you see chat and task.

嗯。

Mhm.

它们开始通过 Argon 给你那种智能体式体验。现在这项能力大幅提速了，而且我们最近在企业活动上宣布，也要把它带给消费者。随着时间推移，很快在 Gemini 里你就能提出请求，而这是一个连续体——你不需要去想自己是在聊天还是在派一个智能体式任务，它会自动做正确的事，并且能在后台安全地持续运行、替你完成任务。从实际数据看，我看 Gemini 应用里进来的请求，大概去年你给它们打标签的话，大多数会是聊天，但已经有稳定增长的比例是 Gemini 里的智能体式任务。

And they start giving you that agentic experience with Argon. Now that capability is dramatically accelerated, and what we announced recently at our enterprise event, we are bringing that to consumers as well. Over time, soon in Gemini you will be able to ask something and it's a continuum — you don't need to think about whether you're asking a chat or an agentic task, it'll just automatically do the right thing, and it can work persistently in the background securely and get things done for you. Already empirically, when I look at what's coming into Gemini in our app, about last year you would have tagged them and most of them would have been chats, but a steadily growing percentage is already agentic tasks in Gemini.

所以是的，我们会把个人智能体带给消费者。它会在 Gemini 应用里工作，在 Google 搜索里工作，在 Gmail 里工作——你可以调用它、喊它等等。但我们想确保把它做对。它将由 Argon 驱动，所以我们非常期待把它带给用户。

So yes, we are going to bring personal agents to our consumers. It'll work in the Gemini app, it'll work in Google Search, it'll work in Gmail — you can address it, call it, etc. But we want to make sure we're getting it right. It'll be powered by Argon, and so we're very excited to bring it to our users.

我有那个 Pro 订阅——我已经被安排上了。所以我很期待多用用。好，几个随机问题。Google 已承诺并向 Anthropic 投入了、承诺了数百亿美元。你在 Dario 身上看到了什么，让你愿意对他做出如此深度的承诺和投资？

I've got that Pro subscription — I've been tasked. So I'm looking forward to using it more. All right, a couple of random questions. Google's committed and invested and has committed tens of billions of dollars to Anthropic. What do you see in Dario that makes you want to have such a deep commitment and investment in him?

很早的时候，我们看 Anthropic 的创始团队，就很清楚这是一支世界级的团队，参与过最早的一些 GPT 模型。所以我们很快就能判断，他们会处在最前沿。我们一直试图识别并投资那些处在技术最前沿的公司。我们在 SpaceX 身上这么做过，在 Anthropic 身上也这么做过。这就是我们看到的。而且我们有机会合作——其中一些合作也延伸到了我们 TPU 的构建上。所以我想这就是我们的思路。

Very early on when we looked at the founding team of Anthropic, it was very clear to us this was a world-class team which was involved in some of the earliest GPT models. So we were quickly able to assess that they would be at the frontier. And so we've always tried to identify and partner with companies which would be at the cutting edge of technology. We've done that with SpaceX. We've done that with Anthropic. That's what we saw. And we had a chance to partner — some of the partnerships turned into the construct of our TPUs as well. So I think that's how we approached it.

好。和 Apple 方面，你们曾有一个合作，把 Gemini 授权给他们，帮助改进 Siri。这个合作是在扩大、收缩，还是原地不动？

Okay. With Apple, you had a partnership with them to license Gemini to them to help improve Siri. Is that partnership expanding, contracting, standing still?

你看，对我们和 Apple 双方来说，这个合作都是关于把 AI 的能力转化为更好的用户体验。我觉得 Apple 在整合方面做得非常出色——他们定制了这些模型并把它构建进 Siri——看到反响这么好真是太好了。所以我们致力于做更多。

Look, for both us and Apple, that partnership was about translating the capabilities of AI into better experiences for users. And I thought Apple has done a phenomenal job of incorporating — they customize these models and build it into Siri — and it's been great to see the reception. So we are committed to doing more.

你们会和他们做更多。

You're going to do more with them.

没有什么具体可以评论的，但这是一个运作良好的合作。

There's nothing specific to comment, but it's a partnership that's working well.

太空数据中心。你们已经宣布要在这方面做更多。到目前为止，关于这个项目的可行性，你们学到了什么？

Data centers in space. You've announced that you're doing more there. What have you learned so far about the feasibility of that project?

这是一个非常令人兴奋的项目。这样的项目让我想起——我至今还记得在沙漠里看 DARPA 大挑战赛的那一刻，看着 Stanley 那辆车参赛，它后来成了 Waymo 的基础。所以对所有这些技术，你都会处在这样一个时刻：它看起来有点像科幻，但 10 到 20 年后，它实际上在以有意义的方式运作。对太空我也有同样的感觉。我们一直相信要在这类技术上大胆下注。量子就是一个很好的例子——我们大概已经走了 10 年。远不止——我们差不多走了 20 年。太空，我们才走了两三年。而我很清楚，这在未来会是非常合理的路径——处在太阳同步轨道上，7×24 小时获取我们能获得的最丰富的能源，对于我们即将需要的算力指数级增长来说，会非常合理。但话虽如此，这是一个有挑战的项目；团队有严格的里程碑。所以把 TPU 送上太空是一个超级令人兴奋的里程碑，就在我们说话的这会儿，我们正期待从它们那里收到一个 hello world 时刻。所以我期待看到它发生，我们也会继续在这里投资。

It's a very exciting project. Projects like this remind me — I still recall that moment in the desert for the DARPA Grand Challenge, watching the Stanley car compete in it, which ended up being the basis for Waymo. And so with all these technologies, you're at a moment in time where it looks a bit like sci-fi, but 10 to 20 years down the line, it's actually working in a meaningful way. I feel the same with space. We've always believed in betting big on these technologies. Quantum has been a great example of it — we're probably 10 years into the journey. Way more — we are almost 20 years into the journey. Space, we are two to three years into the journey. And it's clear to me that it's an approach which will make a lot of sense in the future — being in sun-synchronous orbits with 24/7 access to the most abundant form of energy we have access to will make a lot of sense for this exponential increase in compute that we are going to need. But having said that, this is a challenging project; the team has disciplined milestones. So it was a super exciting milestone to send the TPUs up to space, and sometime as we are speaking we are expecting a hello world moment from them. So I look forward to seeing that happen, and we'll continue investing here too.

你最近——也许是去年——说过，Google 内部写的代码中有 25% 是由 AI 写的。

You recently — was it maybe last year — you said that 25% of code written within Google was written by AI.

今天的数字是多少？

What's the number today?

我是说，我们远远——我觉得我们几乎达到了那个数字。在今天我们的世界里，这个数字某种意义上已经没什么意义了，因为我们现在接触的每一部分代码都深受 AI 影响。比如我们用 Argon 在追踪的东西——我们现在能看到它已经是一个拐点。人们正在用 Argon 去攻克复杂的项目，包括影响数百万行代码的项目。所以自从我们发布 Argon 以来的四周里，我们确实看到了 30% 的销售增长，更重要的是，几乎同样幅度的增长——回滚减少了三分之一。所以我认为人们用它写代码更高效，代码质量也更好。所以我们现在在用更新的指标来追踪。

I mean, we have far — I think we have almost that number. In our world today it's like meaningless in the sense that every part of code we now touch is influenced by AI deeply. What we are tracking with Argon, for example — we are now able to see it's been a point of inflection. People are using Argon to really go tackle complex projects, including projects which affect millions of lines of code. So we have definitely, in the four weeks since we've launched Argon, seen a 30% increase in sales, and more importantly, almost a similar increase — a one-third decrease in rollbacks. And so I think people are writing code more productively and better code with it. So we are tracking it with newer metrics now.

说到对数据中心的反对声音——你怎么看？如果你不能按自己想要的方式在美国建数据中心，你还能去哪里建？

Talking about the opposition to data centers — how do you feel about that? And if you can't build data centers in the US the way that you want to, where else do you go?

我们非常致力于以非常负责任的方式建设数据中心。当我看着我们在一些相处最久的社区里最大的数据中心——我会去走访，和人们交谈——有人会来告诉我们，因为我们在这些社区的投资，他们的孩子上了大学。我们也在这些社区里成长。所以对我来说，关键在于把它做对：尽早与社区接触，透明地去做。我们承诺，在我们引入新增电力的地方，我们会覆盖这些项目 100% 的用电成本。要理解对周边社区的影响，并确保尽早着手解决。所以我认为，只要做得对，你就能继续取得进展。但我们也在不断把反馈纳入考量。我确实担心并非所有人都以同样的方式行事。

We're very committed to building our data centers very responsibly. When I look at our largest data centers in some of the communities where we have been the longest — I visit them, I talk to people — there are people who would come and tell us their kids have gone to college because of our investments in these communities. And we've grown in those communities. So to me this is about doing it right: being engaging with the communities early, doing it transparently. We are committed to making sure where we are bringing incremental power, we cover 100% the cost of our power needs for these projects. It's understanding the impact on neighborhoods and making sure you're working early to address them. So I think done rightly, you can continue making progress. But we are constantly taking feedback into account. I do worry not everyone operates the same way.

是啊，人们对数据中心突然出现在自家后院感到意外。

Yeah, people have been surprised by data centers showing up in their backyard.

所以我认为这些都很重要——作为一个行业，我们应该继续做得更好。但部分原因在于，在 20 年的时间尺度上，我们在考虑太空数据中心这类事情，是为了确保我们能够把这些数据中心放在权衡取舍可能不同的地方。

So I think those are all important — as an industry we should continue to do better. But part of the reason, in a 20-year time frame, we are thinking about things like data centers in space is making sure we are able to put these data centers where maybe the trade-offs will be different.

现在 AI 领域有没有哪门尚未被验证的学科，是 Google 最努力想攻克的？人们谈到过持续学习，就是让这些 AI 在发布之后还能持续更新，而不是每个会话都忘掉。

Is there an unproven discipline in AI right now that Google's working hardest to crack? So people have talked about continuous learning, getting these AIs to continue to update after they get released as opposed to forget through every session.

你看，我们有多个并行的押注。你在 Argon 上看到了，我们有一百万输出 token。早期我们的 Gemini 模型，在输入侧有 100 万上下文窗口。这些都是我们持续在做的突破的例子。你怎么看合成数据？你怎么看持续学习？你怎么看更高效的架构？所以我们确实把相当比例的算力投入到这些研究押注上。而且它总是相辅相成，包括我们的世界模型。所以这些都是我们在下的注。有时作为这些押注的一部分，会冒出一些想法，直接进入主线模型，所以也许你不会完全公开谈，但我认为我们在这些领域有多个押注。

Look, there are multiple parallel bets. You saw with Argon, we have a million output tokens. With our Gemini models early on, we had a 1 million context window on the input side. These are all examples of breakthroughs we are constantly working on. How do you think about synthetic data? How do you think about continual learning? How do you think about more efficient architectures? So we do dedicate a meaningful percentage of our compute to these research bets too. And it always goes hand in hand, including our world models. So those are all bets we are placing. Sometimes as part of these bets there are ideas that emerge which just make their way into the mainline models, and so maybe you don't fully talk about it, but I think we have multiple bets in these areas.

好。我想用这个来收尾。像你这样深入参与科技行业，其中有一种内在的乐观主义，对吧？一种信念：只要你去做，你就会改善世界。这是科技行业根深蒂固的东西。现在，这也有另一面，对吧？它需要押注，往往是押很大的注，希望你在做的东西会以你想要的方式开花结果。但它并不总是那样。我举个例子。我刚去了萨尔瓦多——嗯，我是隔了 5 年去的——我在一个小镇看到人们对 Bitcoin 和区块链有着惊人的热情，有工程在施工，一片繁荣。然后事情并没有像人们预期的那样发生。这次去，那里基本成了鬼城，有很多未完工的建设项目。我觉得要做你做的事，需要一种乐观主义，认为值得去追求，即使押注没有成功，而且可能带来后果。所以我很想听听你的看法——你在斯坦福毕业典礼演讲中谈过这个话题，谈到乐观的必要性。但那是一种有细微差别的乐观，对吧？所以我很想听听你的看法：你是怎么保持这种乐观的，它是纯粹的乐观，还是某种有所保留的乐观？

Okay. I want to end with this. To be involved in the tech industry the way that you are, there's a certain inherent optimism to it, right? A belief that if you build, you will improve the world. And this is something deep within the tech industry. Now, there are sides to that, right? It requires betting, often betting a lot, in hoping that the thing that you're working on is going to come to fruition in the way that you want it to be. But it doesn't always work that way. I'll give you an example. I was just visiting El Salvador — well, I went in a 5-year bracket — and I saw in this one town just an amazing amount of enthusiasm for Bitcoin and the blockchain, and there was construction happening and it was booming. And then it didn't happen the way that folks expected. And this visit was basically a ghost town with a lot of unfinished construction projects. I think to do what you do, it requires a type of optimism that says it's worth going for, even if the bets don't pan out, and there could be consequences there. So I'd just love to hear your perspective — you spoke about this at your Stanford commencement address, about the need for optimism. But it's a nuanced form of optimism, right? So I'd love to hear your perspective on how you maintain that, and is it a pure optimism or is it a couched optimism in a way?

你看，我真心有一种深深的信念：改进技术、改进技术的可及性，会让人们的生活更好。这是我个人的人生经历。所以由此产生了那种深深的信念和乐观。但与此同时，我一直觉得人类的旅程始终是努力驾驭技术，让它为社会良好运转。所以这种乐观不是那种你想建什么就建什么的乐观——不是的，你要努力确保你是在为人类服务而构建技术，让人类境况变得更好，这必须是我们前进时的根本目标。有了这个框架，我非常非常乐观。而且我认为，我们即将在许多领域看到的那种进步，现在还处在非常早期的阶段。我们都在谈医疗健康之类的领域，但我认为很多进步还在前面——能够以根本上更便宜的方式制造许多许多东西，而这些东西长期以来一直在变贵；发现新材料等等，真正扭转曲线。我觉得今天很难一步看清这一切的影响，但它们将会非常深远。所以我很乐观。但我们也必须恰当地担忧这项技术可能出错的所有方式，这样你才能确保解决那些担忧。所以我认为两者是相辅相成的。

Look, I genuinely have a deep conviction that improving technology and improving access to technology makes people's lives better. It's been my personal experience in my life. So out of that comes that deep conviction, optimism. But with that, I've always felt humanity's journey is always about working hard to harness that technology to make it work well for society. So the optimism is not an optimism that you just build whatever you want whenever — no, you work hard to make sure you're building technology in the service of humanity, to make the human condition better, and that's got to be the defining goal as we are approaching it. And with that framework I'm very, very optimistic. And I think we are in very early stages of the kind of progress that we are going to see in many areas. We all talk about areas like healthcare and stuff, but I think a lot of the progress is ahead — the ability to make things in a fundamentally cheaper way on many, many things where things have been getting expensive for a long, long time; discovering new materials, etc., which really bend the curve. I think it's tough to step today and understand the impact of all that, but they're going to be very profound. So I'm very optimistic. But we have to be rightfully concerned about all the ways this technology can go wrong, and that's how you make sure you solve those concerns. So I think they both go hand in hand.

好的，Sundar，我知道你不太做这类访谈。非常感谢你今天抽出时间，很高兴有机会再次见到你、和你交谈。

Well, Sundar, I know you don't do many of these. I really appreciate you spending the time with me today, and it was great to get a chance to see you and speak with you again.

谢谢你，Alex。我很感激。

Thanks, Alex. I appreciate it.

谢谢。好了各位，非常感谢大家的收听和观看，我们下期 Big Technology Podcast 再见。

Thanks. All right everybody. Thanks so much for listening and watching, and we'll see you next time on Big Technology Podcast.