格雷格·布罗克曼谈会自我改进的 AI

我认为非常清楚的是，我们将在未来几年内拥有 AGI，尽管它仍然会有不均衡的表现，但几乎所有使用电脑的智力任务的门槛都会降低。AI 将能够完成这些任务。OpenAI 最可怕的时刻实际上是在我们推出 ChatGPT 之后。我记得在假日派对上，感受到一种‘我们赢了’的氛围。我从未有过这种感觉。我当时想，不，我们是弱者。我们一直都是。从我们推出 ChatGPT 的那一刻起，我记得和我的团队有过这样的对话。我说：‘我们应该买多少算力？’我说：‘全部。’我说：‘不，不，不，真的，我们应该买多少算力？’我说：‘无论我们尝试建造多少，我知道我们无法跟上需求。’

I think it's extremely clear that we are going to have AGI within the next couple of years in a way that it's still going to be jagged, but that the floor of tasks will just be almost for any intellectual task of how you use your computer. The AI will be able to do that. The scariest moment at OpenAI was actually after we launched ChatGPT. And I remember being at the holiday party and just feeling this vibe of 'we won.' I have never felt that. I was like, no. We are the underdog. And we always have been. From the moment we launched ChatGPT, I remember talking with my team having this exact conversation. I said, 'How much compute should we buy?' I said, 'All of it.' I said, 'No, no, no, really. How much compute should we buy?' I said, 'No matter how much we try to build, I know we're not going to be able to keep up with the demand.'

OpenAI 联合创始人兼总裁 Greg Brockman 加入我们，讨论 AI 最有前景的机会，OpenAI 如何利用这些机会，以及超级应用的含义。Greg 今天来到我们的演播室。Greg，很高兴见到你。

OpenAI co-founder and president Greg Brockman joins us to talk about AI's most promising opportunities, how OpenAI plans to capitalize on them, and what the super app is all about. And Greg is with us here in studio today. Greg, great to see you.

谢谢邀请。

Thank you for having me.

我们正在谈论的时机是，OpenAI 正在关闭视频生成，并将精力集中在一个超级应用上，该应用将结合商业和编码用例。从外部来看，包括我自己在内，我们这些观察者觉得 OpenAI 在消费者领域获胜，现在正在转移资源。发生了什么？

Well, we're speaking at a time where OpenAI is shutting down video generation and focusing its energies on a super app, which is going to combine business and coding use cases. And I think from the outside, those of us watching this are like, including myself, OpenAI's winning in consumer, and now it's shifting its resources. What is happening?

嗯，我的看法是，我们一直处于这样一个世界：我们正在开发深度学习技术，看看它是否能产生我们一直设想的那种积极影响。它能否用于构建帮助人们生活的应用。我们还有一个分支，说：‘让我们实际尝试部署这项技术，无论是为了维持业务，还是开始获得一些实际影响的实践，为这项技术真正成熟、成为我们创办这家公司时所想象的一切做准备。’我认为我们现在正处于一个时刻，我们真正看到了这项技术，它将会成功。我们正在从基准测试和近乎智力上的能力展示，转向为了进一步发展它，我们需要在现实世界中看到它，并从人们在知识工作和其他应用中使用它的方式中获得反馈。所以，我认为这是一个更大的战略转变，因为技术所处的阶段。这并不是说我们要从消费者转向 B2B。实际上，我们在说的是，哪些是最重要的应用，我们可以集中精力，因为我们不能什么都做，对吧？但哪些事情我们可以实现，并且它们之间会产生协同效应，带来有意义的影响，帮助提升每个人。当我们列出清单时，有消费者方面，你可以想到很多，但有一个个人助理，对吧？一个了解你、与你的目标一致、帮助你实现生活中任何愿望的东西。还有创意表达、娱乐和许多其他应用。在商业方面，如果你退一步看，它更像是一件事：你有一个困难的任务，AI 能去做吗？它是否有所有上下文来做这些事情？对我们来说，非常清楚的是，优先级列表中有两件事排在首位。一个是个人助理，另一个是能够为你解决难题的 AI。当我们看看我们拥有的算力，我们甚至没有足够的算力来支持这两件事。一旦我们开始添加许多其他应用，许多其他 AI 将非常有用、能帮助人们的事情，我们根本不可能全部做到。所以，我认为这是对技术成熟度及其将很快产生的巨大影响的认识，以及我们需要优先考虑并实际选择我们想要闪耀并真正带给世界的那组应用。

Well, the way I would think about this is that we have been in a world where we're developing this technology, deep learning, to really see can it have the positive impact that we have always pictured. Can it be used to build applications that help people in their lives. And we've separately had an arm that's saying, 'Let's actually try to deploy this technology, whether that's to help sustain the business, to start getting some practice with getting real-world impact, those kinds of things, for the time when this technology actually comes to fruition, that it actually becomes everything that we've imagined, that we started this company to try to have.' And I think that we're at a moment now where we've really seen this technology, it's going to work. And that we're moving out of testing on benchmarks and sort of these almost cerebral demonstrations of capability to it actually being the case that for us to develop it further, we need to see it in the real world and get feedback from how people are using it in knowledge work and various applications. And so, the way I think about it is that this is a bigger strategic shift because of the phase of the technology. And it's not so much that we're saying we're moving from consumer to B2B. It's really what we're saying is that what are the most important applications that we can focus on because we can't focus on everything, right? But what are the things that we can bring to life that will actually synergize together as we build them and that will deliver meaningful impact and help elevate everyone. And when we look at the list, so there's consumer, you can think of it as many things, but there's a personal assistant, right? Something that knows you, that's aligned with your goals, that's going to help you achieve whatever it is that you want in your life. There's also creative expression and entertainment and many other applications. On the business side, maybe you can if you zoom out, it looks more like one thing of just you have a hard task, can AI go do it? Does it have all the context to do all these things? And for us, it's very clear that there the stack rank includes two things at the top. One is the personal assistant, the other is the AI that can go and solve hard problems for you. And when we look at the compute we have, we are not even going to have enough compute to fund those two things. And then once we start adding in many other applications, many other things that AI is going to be very useful for, is going to help people with, we just can't possibly get to all of them. And so, I think that this is a recognition of the maturation of the technology and the incredible impact it's going to have very quickly and our need to prioritize and to actually pick the set of applications that we want to shine and to really bring to the world.

当我听到你谈论 OpenAI 的各种赌注时，你描述的方式之一是，OpenAI 可以成为迪士尼的一个版本，就像迪士尼一样，你有一个核心的引人注目的优势，然后以不同的方式将其分拆出去。所以，迪士尼有米老鼠，然后它可以做电影、主题公园和 Disney Plus。对于 OpenAI 来说，它是模型，你可以做视频生成，成为这个助手，然后帮助企业和工作。那么，是否不再可能拥有那种核心优势，然后以各种方式分拆出去？你是否已经意识到，基本上是该做出选择的时候了？

And when I've heard you talk about OpenAI's various bets, one of the ways that you've described it is that OpenAI can be a version of Disney or a like Disney where you have this core compelling advantage at the center and then you farm it out in different ways. So, Disney has Mickey Mouse and then it can do the movies and the theme park and Disney Plus. And for OpenAI, it's the model and you can do video generation and be this assistant and then help with enterprise and work. So, is it no longer possible then to have that sort of central advantage and then be able to farm it out in all sorts of ways? Like, have you decided Have you come to this realization that basically like it's time to pick or choose?

嗯，我实际上认为在某些方面，这个比喻比以往任何时候都更真实，但重要的是要认识到，从技术上讲，Sora 模型——顺便说一句，它们是令人难以置信的模型——是技术树的一个不同分支，与核心推理 GPT 系列不同。它们只是以非常不同的方式构建的。在某种程度上，我们确实在说，同时追求这两个分支对于这些应用来说非常困难。现在，我们实际上正在机器人技术的背景下继续 Sora 的研究项目，对吧？我认为这显然将是一个变革性的应用，但它仍处于研究阶段，对吧？机器人技术还没有真正成熟，也没有像我们明年将在知识工作中看到的那种技术起飞那样部署。所以，这是认识到，在目前这个时刻，我们真的需要把主要精力放在开发 GPT 系列上，这不仅仅意味着文本，也不仅仅意味着智力上的东西。例如，双向通信，拥有一个出色的语音到语音界面，这也将使这项技术非常可用和有用，但它不是技术树的不同分支。它基本上是一个模型，我们只是以稍微不同的方式调整它，就像你描述的那样。所以，我认为如果你分支太多，拥有两个不同的产物，在算力有限的世界里很难维持。而算力有限的原因是因为需求太大。人们想用我们创建的每个模型做太多事情。

Well, I actually think that in some ways that story is even more true than it's been, but the thing that's important to realize is technologically that the Sora models, which are incredible models, by the way, are a different branch of the tech tree than the core reasoning GPT series. They're just built in a very different way. And to some extent, we're really saying that pursuing both branches is very hard for us to do for these applications. Now, we are actually continuing the Sora research program in the context of robotics, right? Which I think is very clearly going to be a transformative application, which is still a little bit in the research phase, right? That robotics is not really yet mature and deployed in the way that we're going to see this real takeoff of this technology in knowledge work over the next year. And so, it's a recognition of for this moment, we really need to put the primary focus on developing the GPT series and that doesn't just mean text. It doesn't just mean cerebral things. Like, for example, bidirectional communication, having a great speech-to-speech interface, that is something that also is going to make this technology very usable and very useful, but it's not a different branch of the tech tree. It's all kind of one model, and we just sort of tweak that in slightly different ways, kind of like you described. And so, I think there's something about if you branch too far and you have two different artifacts, that is very hard to sustain in a world where there is limited compute. And the reason there's limited compute is because there's so much demand. There's so much people want to do with every single model that we create.

好的，那请谈谈为什么你的赌注不是放在这个世界模型版本上——视频能理解物体去向，这对机器人显然有用。为什么你押注在 GPT 推理模型树，而不是你一直看到 Sora 取得真正进展的这个领域？我的意思是，视频生成从第一代到第三代进步巨大。那么，为什么你的赌注放在那里？

Okay, so talk a little bit then about why your bet is not on this seems like world model version where the video understands where things go. It's obviously useful for robotics. Why is your bet on the GPT reasoning model tree as opposed to this area which you had been seeing real progress with Sora? I mean, to see the progress of video generation, generation 1, 2, 3 was enormous. So, why is your bet where it is?

所以，这个领域的问题是机会太多。对吧？我们在 OpenAI 很早就观察到，我们能想象的一切都有效。现在，不同的想法有不同的摩擦程度、不同的工程工作量、不同的算力需求等等，但每一个不同的想法，只要在数学上合理，你实际上都能开始得到一些相当不错的结果。我认为这展示了底层深度学习技术的力量——真正处理任何问题并抓住核心的能力，让 AI 真正理解生成数据的底层规则。所以，这不是关于数据本身，而是关于理解底层过程，然后能够应用到新的情境中。因此，你可以在世界模型中做到这一点，可以在科学发现中做到，也可以在编码中做到。我认为，当我们考虑这项技术的推广时，一直存在一个争论：文本模型能走多远？文本智能能走多远？能否真正理解世界运作的方式？我认为我们已经明确回答了这个问题——它会走向 AGI。就像我们看到了路线图，而且目前我们已经看到了今年即将推出的更好模型的路线图，OpenAI 内部在决定如何分配算力方面的痛苦，随着时间的推移只增不减。所以，我认为核心可能在于顺序和时机，而在这个时刻，我们一直梦想的应用开始变得触手可及。例如，解决未解的物理问题，对吧？我们最近有一个结果：一位物理学家研究一个问题有一段时间了。他把问题给了我们的模型。12 小时后，我们得到了一个解决方案，他说这是他第一次看到一个模型让他觉得它在思考，感觉这可能是人类永远无法解决的问题，而我们的 AI 解决了它。当你看到这样的事情时，你必须加倍下注，必须三倍下注，因为我们真的可以释放人类的所有潜力。所以，对我来说，这不是这些事情的相对重要性问题。更重要的是 OpenAI 的使命——向世界交付 AGI，我们对它如何惠及每个人的愿景，以及我们有一个技术树，我们知道如何推动它，如何做工程，做进一步的科学研究，然后让它实现。

So, the problem in this field is too much opportunity. Right? The thing we observed very early on in OpenAI is that everything we could imagine works. Now, there's different levels of friction associated with it, different amounts of engineering effort, different compute requirements, all those things, but every single different idea, as long as it's kind of mathematically sound, you actually can start getting some pretty good results. And I think that shows you the power of the underlying technology of deep learning, the ability to really take any sort of problem and to get to the meat of it, to have an AI that really understands the underlying rules that generated the data. So, it's not about the data itself, it's about understanding the underlying process, and then be able to apply to new contexts. So, you can do that in world models, you can do that in scientific discovery, you can do that in coding. And I think that where we are, as we think about the rollout of this technology, is again that there's been this debate of how far will the text models go? How far can text intelligence go? Can you have a real conception of how the world operates? And I think that we have definitively answered that question of it is going to go to AGI. Like we see line of sight and that it is at this point we have line of sight to these much better models that are coming this year and the amount of pain within OpenAI that we've had to decide how to allocate compute, that goes up not down over time. And so I think that maybe the core of it is that we have a it's about sequencing and timing and that in this moment the kinds of applications that we've always dreamed of are starting to come into reach. I, for example, solving unsolved physics problems, right? We had this result recently where a physicist had been working on a problem for some time. He gave it to our model. 12 hours later, we have a solution and he said this is the first time he's seen a model where he felt like he was thinking, that it felt like this is a problem that maybe humanity would never solve and our AI solved it. When you see something like that, you have to double down, you have to triple down because we can really unlock all of this potential for humanity. And so I think for me it's not about relative importance of these things. It's more about what is OpenAI's mission of delivering AGI to the world, our vision of how it can benefit everyone and the fact that we have a tech tree, that we see how to just push it, how to do the engineering, do the further science and research to then have that come to fruition.

好的，我确实想回到你预期的下一代模型，但我想先追问你这一点。今年早些时候，我和 Google DeepMind 的 Demis Hassabis 聊过，有趣的是，他说对他来说最接近 AGI 的是他们拥有的图像生成器 Nano Banana。原因是为了生成图像和视频，它必须理解物体之间的相互作用，并至少对世界如何运作有一些概念。那么，这是否是一个潜在的问题——我的意思是这是一个大赌注，但如果情况如此，OpenAI 加倍押注另一条技术路线是否会错过什么？

Okay, so I do want to come back to the next line of models that you're anticipating, but I want to press you on this for a moment. I was speaking with Demis Hassabis from Google DeepMind earlier this year and interestingly he said that the thing closest that feels closest to AGI for him was Nano Banana, the image generator that they have. And the reason is because for an image generator or a video generator to create the images and the videos that it makes, it does have to understand the interaction between objects and have at least some conception of how the world works. So, is this a potential — I mean it's a big bet, but does OpenAI potentially miss something by doubling down on the other tree if that's the case?

所以，有两个答案。第一，绝对是的。对吧？在这个领域，你仍然必须做出选择，对吧？你必须下注。这实际上就是 OpenAI 的起点——我们真的问过自己：“我们相信通往 AGI 的道路是什么？”然后非常专注地投入。没错，随机向量的和为零，但如果你对齐你的向量，你就能朝一个方向前进。但第二点是，实际上 ImageGen 在 ChatGPT 中非常非常受欢迎，我们也在持续投资和优先考虑它。我们能做到这一点，是因为它实际上并不在世界模型（如扩散模型）的技术分支上，而是基于 GPT 架构。所以，即使数据分布不同，实际的核心技术、核心栈都是一回事。这实际上是 AGI 相当神奇的地方——有时这些看起来非常不同的应用，比如语音到语音、图像生成、文本，而文本本身又是多方面的，比如科学、编码、个人健康信息等，所有这些你都可以在一个技术框架内完成。所以，我从技术角度以及我们公司所关注的，很大程度上是如何尽可能统一我们的努力，因为我们真的认为这项技术将提升和赋能整个经济。整个经济是一个庞大的事物，我们不可能做所有事情，但我们可以尽自己的一份力。

So, two answers. One is absolutely. Yeah. Right, there still is not like in this field you do have to make choices, right? You have to make a bet. And that's actually where OpenAI started is we really said, "What is the path to AGI that we believe in?" And really focused hard on that. Right, the sum of random vectors is zero, but if you align your vectors, then you can go in a direction. But the second point is it's actually ImageGen is something that has been very, very popular within ChatGPT, and that's something we're continuing to invest in, continuing to prioritize. And the reason we're able to do that is because it's not actually on the world model like diffusion model tech branch, it's actually based on the GPT architecture. And so there, even though it's a different data distribution, the actual core technology, the core stack, it's all one thing. And that is actually the pretty wild thing about what AGI is is that sometimes these very different looking applications between speech to speech, image generation, text, and text is by the way itself many faceted of like science and coding and personal like wellness information, those kinds of things. All of that you can do in one technological envelope. And so a lot of what I'm looking at and what we as a company are looking at from a technological perspective is how to have as much unification of our efforts because we really see this technology as being something that's going to uplift and empower the whole economy. The whole economy is a massive thing, and so we can't possibly do all of it, but we can do our part.

这就是人工通用智能中的“通用”部分。

That's the general part in artificial general intelligence.

就是那个 G。

That's the G.

我们说的就是这个。

That's what we're talking about that.

确实如此。

It really is.

说到统一，这个超级应用会是什么？

Speaking of unifying things, what is this super app going to be?

所以，我对超级应用的看法是，它将把编码、浏览器和 ChatGPT 整合在一起。没错。我们想要做的是为你构建一个终端应用，让你真正体验 AGI 的力量，也就是通用性。所以，如果你想想今天的聊天功能，我认为聊天真的会成为你的个人助理，你的个人 AGI，对吧？一个为你着想、了解你很多、与你的目标一致、值得信赖、在数字世界中代表你的 AI。Codex，你可以认为它现在是我们为软件工程师构建的工具，但它正在成为每个人的 Codex。任何想要构建东西的人都可以使用 Codex 让计算机去做他们想做的事情。而且这不再仅仅是关于实际软件。它实际上是关于计算机的使用，比如设置我笔记本电脑上的设置。我忘了怎么做热角？你只需让 Codex 去做，它就做了。对吧？计算机本来就应该是这样的——适应人类，而不是我去适应它们。

So, the way I think about the super app is it's going to bring together coding, browser, and ChatGPT. That's right. So, what we want is to build an endpoint application for you that really lets you experience the power of AGI. So, the generality. And so, if you think about what chat is today, I think chat is really going to become your personal assistant, your personal AGI, right? An AI that's looking out for you, that knows a lot about you, that's aligned with your goals, that's trustworthy, that kind of represents you in this digital world. Codex, you can think of as right now it's been a tool that we built for software engineers, but it's becoming Codex for everyone. That anyone who wants to build can use Codex to get the computer to do the thing that they want. And it's not just about the actual software anymore. It's really about the use of computer, whether it's to set up settings on my laptop. Like I forget how to do the hot corners? You just ask Codex to do it, it just does it. Right? That's what computers were always supposed to be — contort to the human rather than me contort to them.

想象一个应用，你想让电脑做的任何事情，都可以直接问它。AI 内置了浏览器使用功能，可以实际使用网页浏览器，而且你可以监督 AI 在做什么。你所有的对话，无论是聊天、写代码还是通用知识工作，都以统一的方式整合在一起。AI 有记忆，了解你。这就是我们正在构建的东西。但这只是冰山一角，因为这只是表面。对我来说，更重要的是技术上的统一。过去几年里，已经不再仅仅是模型本身，而是关于“套件”：模型如何获取上下文？如何连接到世界？可以采取什么行动？与模型交互的循环如何工作？我们之前有多个不同的实现，现在正在整合它们。我们将有一个统一的版本，最终形成一个 AI 层，可以以很薄的方式指向特定应用。你可以为金融或法律领域构建一个小插件、小技能或小 UI，但通常你不需要，因为会有一个非常广泛的超级应用。

Imagine one application where anything you want your computer to do, you can ask it. There's computer use browsing built in for an AI to use a web browser, and you can oversee what the AI is doing. All your conversations, whether for chat, code, or general knowledge work, are unified in one way. The AI has memory and knows about you. That's what we're building. But it's really an iceberg because that's the tip. What's more important is the technological unification. Over the past couple years, it's no longer just about the model, but about the harness: how does the model get context, how is it connected to the world, what actions can it take, how does the loop of interacting with the model work? We had multiple implementations and we're converging them. We'll have one version of that, an AI layer that can be pointed at specific applications in a thin way. You can build a little plugin, skill, or UI for finance or legal, but generally you won't have to because there'll be one super app that's very broad.

这个应用是针对商业用例还是个人用例？

Is this app for business use cases or personal use cases?

两者都有。就像电脑，比如你的笔记本电脑，它是个人用还是商用？两者都是。它是你的个人机器，为你提供通往数字世界的界面。这就是我们想要构建的。

Both. Just like a computer, like your laptop, is it for personal or business? Both. It's your personal machine that gives you an interface to the digital world. That's what we want to build.

从非商业的角度来看，我在个人生活中使用这个超级应用。我会用它做什么？我的生活会发生什么变化？

From a non-business standpoint, I'm using the super app in my personal life. What am I using it for? How does my life change?

可以把它想象成你现在使用 ChatGPT 的方式。人们用它来做各种令人惊叹的事情：起草婚礼致辞、对某个想法寻求反馈、经营小生意。任何这些问题都应该可以通过超级应用来解决。ChatGPT 一直在进化：它以前没有记忆，就像和一个陌生人对话。如果它能记住交互并访问上下文，比如你的邮件和日历，了解你的偏好，它会更强大。例如，ChatGPT 有一个叫“脉搏”的功能，每天根据对你的了解，推荐你可能感兴趣的内容。超级应用将做到所有这些，而且会更深入、更丰富。

Think of it like how you use ChatGPT now. People use it for diverse amazing applications: drafting a wedding speech, getting feedback on an idea, working on a small business. Any of those questions should be things you can go to the super app for. ChatGPT has been evolving: it used to have no memory, it was like talking to a stranger. It's more powerful if it remembers interactions and has access to context, like your email and calendar, and knows your preferences. For example, ChatGPT has a feature called 'pulse' that surfaces things you might be interested in based on what it knows about you. The super app will do all that, but much deeper and richer.

你们计划什么时候发布？

When are you planning to ship it?

我们将在未来几个月内逐步推进。我们应该会发布完整的愿景，但会分阶段进行。我们从 Codex 应用开始，它集两种功能于一身：一个可以使用工具的通用智能体套件，和一个知道如何编写软件的智能体。这个套件可以用于很多事情：连接到电子表格、Word 文档，帮助进行知识工作。我们将让 Codex 对通用知识工作更加易用，因为我们在 OpenAI 内部已经看到了它的自然采用。这是第一步，后续还有很多。

We're taking incremental steps over the next couple months. We should have shipped the complete vision, but it will come in pieces. We're starting with the Codex app, which is two things in one: a general agent harness that can use tools, and an agent that knows how to write software. That harness can be used for many things: hooked up to spreadsheets, word documents, helping with knowledge work. We'll make Codex more usable for general knowledge work, as we've seen organic adoption within OpenAI. That's the first step, with many to come.

昨天我和你的一位同事聊了 Codex。他提到有人用 Codex 来帮助视频编辑：它构建了一个 Adobe Premiere 插件，开始将视频分成章节，并开始编辑。

I was speaking with one of your colleagues yesterday about Codex. He mentioned someone used Codex to help with video editing: it built a plugin for Adobe Premiere, started separating it into chapters, and started the edit.

我很高兴听到这个。这正是我们希望这个系统能发挥作用的地方。有趣的是，Codex 应用最初是为软件工程师构建的，但对非软件工程师的可用性很低，因为有一些小问题，比如开发者知道如何修复的错误。尽管如此，从未编程过的人也在用它来构建网站、自动化与软件的交互，并获得杠杆效应。我们传播团队的一位成员用它连接 Slack 和邮件来综合反馈。所以我们完成了最困难的部分：一个聪明且有能力的人工智能。现在我们需要做更简单的部分：让它广泛有用，并消除入门障碍。

I love hearing that. That's exactly the kinds of things we want this system to be useful for. It's been interesting seeing the Codex app originally built for software engineers, but its usability for non-software engineers is low because of little things like errors that developers know how to fix. Despite that, people who have never programmed are using it to build websites, automate interactions with software, and get leverage. Someone on our communications team used it hooked up to Slack and email to synthesize feedback. So we did the super hard part of an AI that is smart and capable. Now we have to do the easier part of making it broadly useful and removing barriers to entry.

看看竞争格局，Anthropic 有 Claude 应用：Claude 聊天机器人、Claude 协作写作、Claude 代码。他们有自己的超级应用版本。你认为 Anthropic 看到了什么让他们更早达到这个位置？你们追赶的机会有多大？

Looking at the competitive landscape, Anthropic has the Claude app: Claude chatbot, Claude co-write, Claude code. They have a version of a super app. What do you think Anthropic saw that got them to this position earlier, and what are your chances of catching up?

如果回到 12 到 18 个月前，我们一直专注于编程领域。我们在编程竞赛和脑力活动上一直有最好的数据。但我们没有在最后一英里的可用性上投入那么多。

If you rewind 12-18 months, we always focused on coding as a domain. We always had the best numbers on programming competitions and cerebral things. But we didn't invest as much in that last mile of usability.

我们真正在思考的是：这个 AI 非常聪明，能解决所有编程竞赛难题，但它从未见过真实世界的代码库——那些混乱、不完美的代码。我认为这是我们之前落后的地方。大约去年年中，我们开始认真对待这个问题。我们组建了一个团队，专注于找出所有差距，所有真实世界中存在的、我们尚未遇到的混乱情况。我们如何获取训练数据，构建训练环境，让 AI 体验真正的软件工程是什么样子，比如被奇怪的方式打断等等。现在我可以说到这个点上我们已经赶上了。当人们直接对比我们和竞争对手时，他们往往更喜欢我们。我们知道在前端方面还有差距，我们会修复它，但这是我们一直以来的总体方向：思考产品的端到端可用性，而不仅仅是模型，然后单独构建一个东西，对吧？真正把它看作一个整体产品。我们在做研究时，会思考它将如何被使用。这是我们在 OpenAI 内部一直在推动的改变。所以，我认为看待这件事的方式是，我们有非常棒的模型升级即将到来。今年我看路线图，真的令人振奋。未来可能实现的事情。然后我们现在真正专注于让最后一英里的可用性也跟上。

Of really trying to think about, okay, this AI is so smart, it can solve all these great programming competitions, but it's never seen someone's real-world codebase, which is messy and not quite as pristine as the world that it has experienced. And I think that is something that we were behind on. But about maybe mid last year is when we got very serious about that. And we had a team very focused on what are all the gaps, what are all the kind of messiness the real world would have and we haven't encountered. How do we actually get training data that build training environments that let the AI experience what it's like to actually do software engineering, be interrupted in weird ways, all those things. And I'd say at this point we are caught up. When people go head-to-head, for us versus competitors, the people tend to prefer us. We do know where we're lagging in front end, we're going to fix that, but this is the general motion that we've been taking is to say that usability of thinking about the product end-to-end, not just a model, and then build a separate thing, right? Really think of it as one product. When we're doing the research, we're thinking about how will it be used. That has been a motion that we've been changing within OpenAI. And so, I think that the way I would look at it is that we have incredible step-up models coming. Like this whole year I look at the road map, it's truly inspiring. What will be possible. And then we've been really focusing now on let's also get the last mile usability.

所以自 2022 年以来，OpenAI 一直是无争议的领导者，但现在竞争显然很激烈。你刚才用了“我们赶上了”这个说法。公司内部氛围是否有所不同？现在不再是像 ChatGPT 那样遥遥领先，而是真正在战斗？我看到一些关于公司内部的报道，说 OpenAI 不再有会议，不再有支线任务，一切都聚焦于此。环境怎么样？氛围变了吗？

So since 2022, OpenAI has been like the undisputed leader and obviously now the competition is intense. Like you just used the phrase we're caught up. Is there a different vibe within the company where it's like now instead of the one that's like far ahead on something like ChatGPT in a real fight? I mean, you're seeing it come out of some of the reporting on what's happening within the company, the fact that there are no more meetings, there's no more side quests at OpenAI. It's all focus on this. How's the environment? Has the vibe changed here?

对我来说，OpenAI 最可怕的时刻实际上是在我们推出 ChatGPT 之后。我记得在假日派对上，我感受到一种“我们赢了”的氛围。我从未有过这种感觉。我想，不，我们是弱者，而且一直是，对吧？这个领域的竞争对手，那些成熟的公司，拥有更多的资本、人力资源、数据等等。OpenAI 凭什么能竞争？在某种程度上，答案仅仅是因为我们从不自满，我们始终觉得自己是挑战者。实际上，看到我们在市场上开始看到这一点，看到其他竞争对手出现并做得很好，对我来说是一件非常健康的事情。在我看来，你永远不能只盯着竞争对手。如果你专注于他们所在的位置，你就会停留在那里，而他们已经向前移动了。我认为这正是相反方向发生的事情，对吧？很多人专注于我们所在的位置，而我们得以继续前进。我认为这几乎给了我们这种一致性和公司的统一。我描述过我们曾经几乎把研究和部署看作分开的事情，而现在我们真的想整合它们。这对我来说是一件非常美妙的事情。所以，我认为我们处于这样一个世界：你永远不会像别人说的那么好，也永远不会像别人说的那么差。我觉得一切都很稳定。在模型生产的核心方面，我对我们的路线图和研究投资非常有信心。在产品方面，我们有如此巨大的能量，所有这些汇聚在一起，将成果带给世界。

Well, I would say that for me personally, yeah, the scariest moment at OpenAI was actually after we launched ChatGPT. And I remember being at the holiday party and just feeling this vibe of we won. I have never felt that. I was like, no, we are the underdog and we always have been, right? The competitors in this space, established companies that have just sort of much more capital, much more human resources, data, the whole thing. Why is OpenAI able to compete at all? And to some extent, the answer is only because we never feel complacent, where we always feel like we are the challenger. And it actually for me has been a very healthy thing to see us start to see that in the marketplace, to see other competitors emerge and do a good job. And that is, in my mind, you can never fix on your competitors. If you focus on where they are, then you'll be where they are and they'll already have moved. And I think that that's what's been happening in the other direction, right? A lot of people have focused on exactly where we are and we get to move. And I think that it almost gives us this alignment and unification of the company. And I kind of described how we almost thought of research and deployment as separate things. And now we really want to integrate them. Like, that to me is such a wonderful thing. And so, I'd say that the world that we're in is one where I've never felt like we were you know, you're never as good as they say you are, you're never as bad as they say you are. I think it's just been very steady. And that the core of the model production, that is something where I actually feel extremely confident in our road map, the research investments we've been making. And I think on the product side, we have such great energy that's all coming together to deliver this to the world.

你之前几次暗示你们有一些好模型即将推出。Spud 是什么？有消息说你们完成了 Spud 的预训练，OpenAI 的 CEO Sam Altman 告诉员工，他们应该期待在几周内看到一个非常强大的模型。这是几周前的事。团队相信它真的能加速经济发展，事情进展比我们许多人预期的要快。那么，Spud 是什么？

You foreshadowed a couple times already that you have some good models on the way. What is Spud? The information said that you finished pre-training Spud, and Sam Altman, the CEO of OpenAI, has told the staff that they should expect to have a very strong model in a few weeks. This was a few weeks ago. And the team believes it can really accelerate the economy, and things are moving faster than many of us expected. So, what's Spud?

这是一个好模型。但我认为真正重要的不是任何一个模型。我们的开发流程是这样的：先进行预训练，产生一个新的基础模型，然后在这个基础上进行进一步的改进。这始终是公司许多人共同努力的巨大工程，过去 18 个月我大部分精力都投入在 GPU 基础设施上，支持团队进行训练框架的扩展，以进行这些大规模运行。但之后还有强化学习过程。你让这个 AI 学习了很多关于世界的知识，它应用这些知识，然后我们进行后训练过程，真正说：‘好了，现在你知道如何解决问题了。你在所有这些不同的情境中练习。’然后是行为和可用性的最后一英里。所以，我把 Spud 看作一个新的基础，一个新的预训练模型，我们大约有两年的研究成果在这个模型中得以实现。这将非常令人兴奋。我认为世界将体验到的是能力的提升。对我来说，这从来不是关于任何一个版本的发布，因为一旦我们发布这个版本，它只是我们未来成果的早期版本。我们会在改进过程的每一步做更多工作。所以，我认为我们前进的方向是拥有一个越来越快的进步引擎。而这一步只是其中的一步。

It's a good model. But I think that it's really not about any one model. The way our development process works is you have pre-training, so you produce a new base model that then is the foundation that we build further improvements on top of. And that is always a huge effort across many people in the company, and that's where I've actually been spending most of my efforts over the past 18 months has been really focused on our GPU infrastructure, on supporting the teams that do all of the training frameworks to scale up at these big runs. But then there's a reinforcement learning process. So, you take this AI that has learned lots of things about the world, and it applies that knowledge, and then we do a post-training process where you really say, 'Okay, now you know how to solve problems. You practice it in all these different contexts.' And then here's kind of the last mile of behavior and usability. So, I think of Spud as a new base, as a new pre-train, and that we have had this I'd say it's like we have maybe 2 years' worth of research that is coming to fruition in this model. It's going to be very exciting. And I think that the way that the world will experience it is just improved capabilities. And that for me, it's never about any one release, because as soon as we have this one release, it'll be an early version of what we have coming. We'll do much more of each of these steps of the improvement process. And so, I think that where we're going is this almost just we have this engine of progress that just moves faster and faster. And this one is just one step along the way.

那么，你认为它能做到哪些今天的模型做不到的事情？

And so, what do you think it'll be able to do that today's models can't?

我认为它将能够解决更困难的问题。它会更加细致入微。它能更好地理解指令，更好地理解上下文。人们常说的“大模型气味”就是指当这些模型实际上更聪明、更有能力时，它们会更顺从你。你能感觉到，对吧？当你问一个问题而 AI 没有完全理解时，总是很令人失望。

I think it's going to be able to solve both much harder problems. I think it will be much more nuanced. It'll understand instructions better. It'll understand the context much better. There's this thing called big model smell that people talk about where it's just like there's something about when these models are just actually much smarter, much more capable, that they bend to you much more. You feel it. Right? When you ask a question and the AI doesn't quite get it, it's always so disappointing.

对吧？当你不得不解释时，你会想，‘你真的应该能自己搞明白。’所以，我认为在某种程度上，会有定性的变化，但也有很多定量的变化。对吧？定性上，会出现新的事情，以前你会感到沮丧，从不用 AI，现在你几乎不假思索就用上了。我认为这就是我们将在各个领域看到的情况。我非常兴奋地看到它如何提高上限。对吧？我们已经看到了这些物理应用之类的东西。我认为我们将能够解决更多开放性问题，时间跨度也更长。同时，我也很兴奋地看到它如何提高下限，让你想做的任何事情都变得更有用。

Right? When you have to explain and you're just like, 'You really should be able to figure this out.' And so, I would just think of it as in some ways just qualitatively, there will be but quantitatively lots of shifts. Right? And qualitatively, there will just be new things where you would be frustrated before, you never use an AI for it, and now you just use it without thinking very much. And I think that is what we're going to see across the board. I'm super excited to see how it raises the ceiling. Right? We've already seen these physics applications, things like that. And I think we will be able to just solve way more open-ended problems, way longer time horizons. And then also very excited to see how it raises the floor, where just for anything you want to do, it's just so much more useful for you.

对于日常用户来说，真正感受到变化可能有点难。比如在 GPT-5 发布之前有很多炒作，然后它发布了，实际上公众的初始反应有些失望。但后来，我认为人们意识到它在某些任务上确实很好。或者对于下一系列模型，你预计它会在某些职业中真正被感受到，还是认为它会成为每个人都能广泛感受到的切实改进？

It can be kind of tough for everyday users to really feel the change. Like there was talk about a lot of build-up before GPT-5 came out, and then it came out, and actually the initial reaction was somewhat disappointment among the public. But then, I think people realized that for certain tasks it was really good. Or with these next series of models, do you expect that it'll really be felt in the trenches in certain occupations, or do you think it'll be a broadly tangible improvement for everyone?

我认为这将是一个类似的故事：当你发布它时，会有人尝试它并说，‘这和我见过的任何东西都天差地别。’然后会有一些应用，我们之前并不一定受限于智能。所以如果你有一个更智能的模型，也许你不会立刻感受到。但我认为随着时间的推移你会感受到，因为根本的变化是你对系统的依赖程度？就像你思考我们与 AI 互动的方式，我们对自己认为它能做什么有一个心理模型。而这个心理模型实际上变化相当缓慢。对吧？随着你获得更多经验，它会为你做一些神奇的事情。你会说，‘哦，哇，它能做到？我从未想象过。’例如，我们在获取健康信息等应用中看到了这一点，对吧？我们看到人们，比如我有一个朋友，他用 ChatGPT 来了解他癌症的不同治疗方法。医生告诉他他已经是晚期，无能为力。他实际上用 ChatGPT 研究了很多不同的想法，并因此得到了治疗。这需要你有一定程度的信念，认为 AI 在那个应用中有帮助，你才会真正付出努力从机器中得到一些东西。我认为我们将看到的是，对于任何这样的应用，AI 能帮助你这一点对每个人来说都会变得更加明显。所以我认为这既是技术的进步，也是我们对技术理解的转变和追赶。你会更加依赖它。

I think that it will be a similar story where when you release it, there will be people who will try it and be like, 'This is a night and day different than anything I've seen.' And then there will be some applications where we weren't necessarily intelligence bottlenecked. And so if you have a model that's more intelligent, maybe you won't feel it right there. But I think over time that you will feel it because the fundamental thing that shifts is how much do you rely on the system? Like if you think about the way we all interact with AI, we have some mental model for what we think it can do. And that mental model shifts actually fairly slowly. Right? As you get more experience, it does something magical for you. You're like, 'Oh, wow, it can do that? I never imagined that.' And we see this for example in applications like access to health information, right? That we see people who, you know, like I have a friend who used ChatGPT to understand different treatments for his cancer. And he was told by doctors that he was terminal, that there's nothing they could do for him. He used ChatGPT to actually research a bunch of different ideas and he was able to get treatment that way. And that's something where you need to have some level of belief that the AI is going to be helpful in that application for you to really put in the effort to get something out of the machine. And I think what we're going to see is that for any application like that, it's going to become so much more evident to everyone that the AI can help you. And so I think it's a little bit of the technology getting better, but it's also our understanding of the technology shifting and catching up to that. And you'll be relying on it more.

在 OpenAI 内部，你有一个正在开发中的自动化 AI 研究员，预计今年秋天推出。那是什么？

Inside OpenAI, you have an automated AI researcher in the works that's supposed to come out this fall. What is that?

所以，目前的发展方向是，我们正处于这项技术起飞阶段的早期。起飞意味着什么？起飞是指 AI 在指数级上变得越来越好。部分原因是我们可以用 AI 来改进 AI。所以，我们的开发过程加速了。但我也认为，当我想到起飞时，它也关乎现实世界的影响。在某种程度上，每项技术都是一条 S 曲线。或者如果你放大看，一些 S 曲线最终会变成指数级。我认为这就是我们现在遇到的情况。所以，技术发展正在以越来越快的速度前进，这是一个正在积聚动力的引擎。但在世界上，也有所有这些顺风，因为芯片开发者正在为他们的项目获得更多资源。还有一群人在其上构建，试图弄清楚它如何适应每个不同的应用。所有这些能量都在不断积累，进入 AI 从某种配角变成经济增长主要驱动力的起飞阶段。我认为这不仅仅关乎我们在这些围墙内做的事情。它关乎整个世界、整个经济如何团结起来，共同推动这项技术及其有用性。

So, the direction of travel right now, we are in this early phase of takeoff of this technology. What does takeoff mean? Takeoff is as the AI gets better and better on this exponential. And in part because we can use the AI to make the AI better. So, our development process speeds up. But I also think when I think of takeoff, it's also about real world impact. And in some ways we've been, you know, every technology is an S curve. Or if you zoom out, some S curves that end up being an exponential. And I think that that's what we're encountering right now. So, it's the technology development is moving with increasing speed and it's this engine that's picking up momentum. But it's also in the world, there's all of these tailwinds because there's chip developers that are getting more resourcing into their programs. There's this economy of people who are building on top of it, trying to figure out how it fits into every different application. And all of that energy is just accumulating more and more into this takeoff phase of the AI becoming a kind of sideshow to being the main driver of economic growth. And I think that that is something that it's not just about what we're doing in these walls. It's about how the whole world, the whole economy comes together and are to push forward this technology and its usefulness together.

那么，这个研究员具体会做什么呢？

And the researcher, well, then what will it do exactly?

此外，研究员将是一个时刻，我们正在构建的 AI 现在正在承担更大比例的任务，我们应该能够让它自主运行。我认为关于这意味着什么有很多思考。这并不一定意味着我们只是让它自己运行，然后稍后再回来看它是否做了好事。我认为我们将非常参与管理它。对吧？就像现在，如果你有一个初级研究员，如果你让他们自己待太久，他们很可能会走一条不太有用的路。但如果你有一个高级研究员或有远见的人，他们甚至不一定需要知道机械技能，他们能够提供反馈，审查这个人产生的图表，并根据愿景提供方向，即我希望你完成什么。所以我认为这是一个我们将构建的系统，它将极大地加速我们生产模型的能力，实现新的研究突破，使这些模型在现实世界中更有用、更可用。

Also, the researcher will be a moment where the AI which we're building that right now it's taking a larger percentage of tasks, that we should be able to let it run autonomously. And that I think there's a lot of thought that goes into what that means. And that it doesn't necessarily mean that we just let it off on its own and then come back later and see if it did something good. I think that we are going to be very involved in managing it. Right? Just like right now if you have a junior researcher, if you leave them on their own too long, they're probably going to go down a path that's not very useful. But if you have a senior researcher or someone who has a vision, they don't even necessarily need to know the mechanical skills, they will be able to provide feedback, review the plots that this person's producing and to provide direction in terms of the vision of what is it that I want you to accomplish. And so I think of this as a system that we're going to build that will massively accelerate our ability to produce models, to make new research breakthroughs happen, to be able to make these models more useful and usable in the real world.

对。

Right.

并且以越来越快的速度做到这一点。

And to do that at increasing speed.

所以，抱歉。它会做什么？你是说它会去寻找 AGI，然后尝试？

So, sorry. What's it going to do? Are you going to say go find AGI and it will just try to?

我认为我思考它的方式，大致就是这样。

I think the way I think of it is something like that to first order.

好的。

Okay.

在实际层面，我认为我会把它看作是我们研究科学家所做的完整端到端工作，并能够在硅基中完成。

And at a practical level, I think I would view it as taking the full end-to-end of what one of our research scientists does and be able to do that in silicon.

另一种思考起飞的方式是，AI 的进展从增量式变为积聚动力，然后某种不可阻挡地走向比人类更聪明的智能。你是否担心，就像在那方面有事情顺利的可能性一样，那个过程也有可能出错？

Another way to think about takeoff is their progress in AI goes from incremental to gathering momentum and then sort of this unstoppable march to an intelligence that's smarter than humans. Do you worry that just as there are possibilities for things to go right on that front, there are also possibilities for that progress to go wrong?

我的意思是，我认为绝对是的。我认为获得这项技术好处的方法也是真正考虑风险。

I mean, I think that that's absolutely yes. I think that the way to get the benefits of this technology is also to really think about the risks.

嗯。

Mhm.

如果你从技术角度看我们如何应对技术发展，我们在安全方面投入了很多。一个很好的例子是提示注入。如果你要有一个非常聪明、非常能干、连接了大量工具的 AI，你要确保它不会被别人用奇怪的指令破坏。这是我们投入了很多的领域，我认为取得了非常了不起的成果，有一个很棒的团队在做。有趣的是，有些问题可以类比人类。人类也容易受到钓鱼攻击，以不同方式被欺骗，并不完全理解他们工作的全部背景。我们把那些类比带入开发过程，每当我们发布模型、开发模型时都会思考：我们如何确保它能与人类对齐，真正有帮助。这是我们非常关心的事。我认为还有更大的问题，关于世界、经济、一切如何变化、每个人如何从这项技术中受益。它们不纯粹是技术问题，不纯粹是 OpenAI 自己能解决的。但没错，我思考很多的不只是推动技术前进，还有如何确保我们实现其潜在的积极影响。

And if you look at how we've approached technology development from a technical perspective, we invest a lot in safety, security. A good example of this is prompt injections. If you're going to have an AI that is very smart, very capable, hooked up to lots of tools, you want to make sure that it can't be subverted by someone giving it a weird instruction. That's something we've invested in quite a lot and I think have really incredible results, have an incredible team working on. It's interesting to think about some of these problems where you can make analogies to humans. Humans are also susceptible to phishing attacks, to being deceived in different ways, to not really understanding the full context of what they're working on. We bring those analogies into our development process and think about this whenever we release a model, develop a model: how do we ensure that it's going to be aligned with people and be able to actually be helpful. That is something we care quite a lot about. I think that there are bigger questions about the world, the economy, how does everything change, how does everyone benefit from this technology. They're not purely technical, not purely something that OpenAI on our own will be able to solve. But yes, I think quite a lot about not just pushing forward the technology, but also really about how do we ensure that we have the positive impact that is its potential.

但担忧在于这是一场竞赛，OpenAI 总部内部做的事情也被许多开源玩家复制，他们在安全方面的边界、障碍和保护要少得多。而且如果你说这需要很多人做对很多事情才能有创造力，而一个心怀恶意的人就能造成破坏。至少对我来说，担忧就在于这显然是一场竞赛，进展很快。你的许多同行说过，如果大家都同意停止，我们就停止，但这似乎并不会减缓它。

The worry though is that this is a race and what's being done within these walls at OpenAI headquarters is also being copied by many of the open source players, which have much less bare boundaries and barriers and protection on the safety side of things. And if you said this months that it takes, you know, people getting a lot of things right to be creative and sort of one person with bad intent to be destructive. That's sort of where the concern lies for me at least is just when this is clearly a race, it's going fast. Many of your counterparts have said if everybody agrees to stop it, we'll stop it and it doesn't seem like it's going to slow it all.

所以基本上，回报值得冒险吗？

So is the reward worth the risk basically?

我认为回报值得冒险，但我觉得这在某种意义上太粗略了。我的思考方式是，我们从 OpenAI 成立之初就问：一个美好的未来是什么样的？这项技术如何才能真正提升每个人？你可以认为几乎有两个不同的角度。一个是集中化的观点，认为让这项技术安全的方法是只有一个参与者来构建它，这样你就没有任何压力。你可以真正考虑把它做对，然后等它准备好了再推广给所有人。这在某些方面很难接受，我认为有很多属性你可以考虑用不同的方式来处理，我们更倾向于称之为韧性。把它看作一个开放系统，有很多参与者都在开发这项技术。这不只是关于技术，而是关于建立社会基础设施，帮助这项技术真正顺利发展。如果你想想电力是如何发展的，那是很多人生产的东西，它实际上有危险和风险，但我们也以多种不同的方式建立了安全基础设施，围绕电力的安全标准，不同的利用方式，如何规模化，当达到大规模时有法规，很多人能够以民主化的方式使用，还有检查员。整个系统是围绕那项技术的需求、那项特定技术的特性建立起来的。我认为我们在 AI 中真正看到的一件事是，我们需要广泛的对话。如果这项技术要来改变每个人的一切，人们需要参与其中。它不能由某个集中化的团体秘密进行。所以对我来说，这是这项技术应该如何发展的一个核心问题，我们真正相信的是应该围绕这项技术的发展出现一个韧性生态系统。

I think the reward is worth the risk, but I think that is too coarse-grained of an answer in some sense. The way I think about it is that we've asked from the beginning of OpenAI how what does a great future look like? How can this technology really be something that uplifts everyone? You can think of there almost being two different angles. One is the centralization view of saying that the way to make this technology safe is that you have only one actor building it, so then you don't have any pressures. You can really think about getting it right and then figure out how to roll it out to everyone when it's ready. That's a pretty tough pill in some ways and I think there's a lot of properties that you can think instead about approaching differently, which we prefer to as resilience. To think of it as this open system where there's lots of players who are developing the technology. It's not just about the technology, it's about building societal infrastructure that helps this technology really go well. If you think about how electricity has developed, that's something where lots of people produce it, it actually has dangers and risks, but we also build our safety infrastructure in a diversity of different ways around safety standards for electricity, around different ways of harnessing it, about how you scale it, that there's regulations when you're at these massive scales that lots of people are able to use in a democratized fashion, there's inspectors. There's a whole system that's been built around the needs of that technology, the proclivities of that specific technology. I think that one thing that we have really seen with AI is that it is something where we need this broad conversation. We need lots of people to be aware if this technology is going to come and change everything for everyone, people need to participate in that. It can't be something that's done often secret by just one centralized group. So this has been to me a very core question to how this technology should play out and something we really believe in is this resilience ecosystem that should emerge around the development of this technology.

所以你说我们正处于起飞阶段，在起飞过程中，我想全人类都在经历这个。英伟达 CEO 黄仁勋最近说他认为 AGI 已经实现了。你同意吗？

So you said we're in takeoff, in the middle of a takeoff process and we I guess all of humanity are experiencing this. Nvidia CEO Jensen Huang said recently that he believes that AGI has been achieved. Do you agree?

我认为 AGI 对很多人有不同的定义，很多人会说我们现在拥有的就是 AGI。我觉得可以辩论。但有趣的是，AGI，就像我们现在拥有的技术，非常参差不齐。它在很多任务上绝对超人类。在写代码之类的事情上，AI 就能做到。它确实消除了很多创造事物的摩擦。但有一些非常基础的任务，人类能做，而我们的 AI 仍然吃力。所以几乎可以说，你在哪里划界线？目前这更像是一种感觉，而不是科学。所以对我自己来说，我们肯定在经历那个时刻。如果你 5 年前给我看今天的系统，我肯定会说，‘哦，是的。这就是我们说的。’但它就是不一样。它和我们曾经想象的任何东西都如此不同。所以我认为我们需要适当调整我们的思维模型。所以，你还没到那一步。我想我会说我基本上到了 70%、80%。所以我认为我们非常接近了。而且我认为非常清楚的是，我们将在未来几年内拥有 AGI，它仍然会是参差不齐的，但任务的下限将几乎涵盖任何智力任务，比如如何使用电脑，AI 都能做到。我认为现在我得给出一个有点不确定的答案，因为有点不确定性原理之类的东西。你可以辩论。就我个人定义而言，我认为我们几乎到了，再稍微多一点，我们就绝对到了。

I think that AGI has a different definition to many people and I think that there are many people who would say that what we have right now is AGI. I think you can debate it. But I think that maybe the thing that's interesting is that AGI, like the technology we have right now, is very jagged. It is absolutely superhuman at many tasks. When it comes to writing code, those kinds of things, AI can just do it. And it really removes a lot of the friction to creating things. But there's some very basic tasks that a human can do that our AI still struggle with. So it's almost to say that where do you draw the cutline? It's a little bit more of a vibe than a feeling than it is science at the moment. So I think for myself, we're definitely going through that moment. And if you were to show me 5 years ago the systems we have today, I'd have to go, 'Oh, yeah. That's what we're talking about.' But it's just different. It's so different from anything we ever pictured. So I think we need to adjust our mental models appropriately. So, you're not there yet. I think I'd say I'm basically like 70, 80% there. So I think we're quite close. And I think it's extremely clear that we are going to have AGI within the next couple of years in a way that is still going to be jagged, but that the floor of task will just be almost for any intellectual task of how you use your computer, that the AI will be able to do that. And I think that right now I have to give a little bit of an uncertain answer because there's some it's almost like an uncertainty principle kind of thing. You can debate it. For my own personal definition, I think we're almost there, and with maybe a little bit more, we will absolutely be.

好的。我们要休息一下了，但在休息之前，我想让在家观看的朋友们知道，你和我将在 6 月 18 日再次在旧金山的 SFJAZZ 进行对话。我会在节目说明中放一些信息，如果你想加入那场对话，我希望你报名。好了。

Okay. Well, we're going to go to a break, but as long as we're on the way to the break, I want to let folks watching at home know that you and I are going to be talking again June 18th here in San Francisco at SFJAZZ. So, I will put some information if you want to come join that conversation in the show notes, and I do hope you sign up. All right.

我们马上回来。欢迎回到 Big Technology Podcast，今天我们请到了 OpenAI 联合创始人兼总裁 Greg Brockman。Greg，我想问你，2025 年 12 月发生了什么？因为那似乎是一个转折点，让机器连续编码数小时的想法从理论变成了一个时刻，每个人都觉得‘我想我可以信任它继续工作一段时间了’。那么，到底发生了什么？

We'll be back right after this. And we're back here on Big Technology Podcast with OpenAI co-founder and president Greg Brockman. Greg, let me just ask you, what happened in December 2025? Because it seems like it was an inflection point where all this idea of letting the machine code for hours uninterrupted went from theory to a moment where everyone said, 'I think I can trust this to keep going for a while.' So, what exactly happened?

新模型发布后，AI 从能完成你大约 20% 的任务提升到了 80%。这是一个巨大的转变，因为它从‘嗯，这是个好东西’变成了‘你绝对需要围绕这些 AI 重新调整你的工作流程’。对我自己来说，我确实有这样一个时刻：我有一个用了多年的测试提示——为我建一个网站。我实际上在学编程时建过这个网站，花了几个月。在 2025 年期间，通常需要 4 个小时，用一堆不同的提示才能搞定。但在 12 月，一次搞定。我只问了一次 AI，它就生成了，而且做得很好。

New model releases really went from the AI being able to do like 20% of your tasks to 80%. And that was this massive shift because it went from being kind of a 'yeah, it's a nice thing to do' to 'you absolutely need to retool your workflow around these AIs.' For myself, I very much had this moment where I have a test prompt that I've been using for years: build a website for me. I'd actually built this website back when I was learning to code, it took me months. Used to be over the course of 2025 that it would take like 4 hours, a bunch of different prompts to get it right. In December, one shot. Just asked the AI one time and it produced it and did a great job.

那么，这些模型是如何实现飞跃的？

So, how did those models make the leap?

很大程度上是因为更好的基础模型。OpenAI 的一个特点是，我们一直在改进预训练技术，已经有一段时间了，而在那个时刻，我们得以一窥今年剩余时间将要发生的事情。但这也不是单一因素。我们一直在创新的每一个维度上推进。这些模型的有趣之处在于，在某些方面你会看到这些飞跃，而在某些方面它又是连续的。它不是从 0% 到 80%，而是从 20% 到 80%。所以，从某种意义上说，它只是变得更好了。我认为我们实际上在每个小版本发布中都看到了这种改进。比如在 5.2 和 5.3 之间，我的一位密切合作的工程师，从无法让 AI 完成他做的底层硬核系统工程师工作，到 AI 完全胜任。他给 AI 一个设计文档，AI 就能实际实现它，添加指标、可观测性，运行分析器，改进到正是他希望产出的东西。所以我认为思考方式是，它几乎是‘慢慢地、慢慢地、慢慢地，然后一下子’。但这一切都通过当前有效的东西体现出来。当然，在一年内，有时更早，它会变得极其可靠。

Well, a lot of it is about the better base models. The one thing about OpenAI is that we've been working on improving our pre-training technology for quite some time and in that moment we got to see a little taste of what is going to be coming for the rest of this year. But it's also really about not any one thing. It's about we're constantly pushing on every single axis of innovation. And the thing that's very interesting about these models is in some ways you get these leaps. In some ways it's all continuous. It didn't go from 0% to 80%. It went from 20% to 80%. And so, in some ways it just got better. I think that we've actually seen this improvement continue with every single point release that we've had. Like between 5.2 and 5.3, one of my engineers I work with very closely went from he couldn't get it to do the low-level hardcore systems engineer he does to it absolutely being graded. He gives it a design doc and actually implements it, adds metrics, observability, runs the profiler, improves it to the point that it's the exact thing that he was hoping to produce. And so I think that the way to think about it is it's almost a sort of slowly slowly slowly all at once. But it is all indicated by what's kind of working right now. Certainly within a year, sometimes much sooner, is going to be incredibly reliable.

这让你感到惊讶，因为不久前我在一次采访中听到你谈到 Codex，这个自主编码器，只是为软件开发者准备的。而在这段对话早些时候，你说实际上每个人都能用这个东西。是什么导致你改变了看法？

And it surprised you because I heard you talking on an interview not long ago about how Codex, this autonomous coder, was just for software developers. And earlier this conversation you said actually everyone can use this stuff. What led to the fact that you sort of changed your perspective on that?

嗯，我想我之前一直关注 Codex，它里面有代码，对吧？所以它确实是给程序员用的。考虑到 OpenAI 内部的人，因为我们很多人都是为自己构建软件的工程师，很自然会这么想。但随着这项技术的进步，我们开始意识到，我们生产的底层技术大多与代码无关。它主要是关于解决问题。主要是关于管理上下文和工具，思考 AI 应该如何整合并完成工作。而这使得即使是代码，突然之间任何人都能使用，因为你可以管理一个将要去做工作的东西，对吧？如果你有一个愿景，有想完成的事情，你可以描述你的意图，AI 就能执行并完成。但随后它也开始思考，‘为什么我只关注编码？’比如有很多与 Excel 电子表格、演示文稿相关的非常机械的技能。如果 AI 拥有上下文，它现在拥有原始智能，能够高水平地完成这些事情。所以，如果我们能让它更容易被使用，突然就从‘Codex 是为程序员准备的’变成了‘Codex 是为所有人准备的’。

Well, I think I'd been focusing on Codex and it's got the code in it, right? So it's really been for coders. And thinking about people within OpenAI because many of us are software engineers building for ourselves, it's very natural to think that way. But as this technology has been progressing, we've started to realize that the underlying technology we produced is mostly not about code at all. It's mostly about solving problems. It's mostly about being able to manage contexts and harnesses and think about how an AI should integrate and do work. And that's something that becomes both even for code suddenly anyone can have access because you can manage something that's going to go do work, right? If you have a vision, you have something you want to accomplish, you can describe your intent, the AI can execute, can get that done. But then it also starts to go, 'Well, why am I just focused on coding?' Like there's so much just very mechanical skill associated with Excel spreadsheets, with presentations. And if the AI has the context, it has the raw intelligence now to be able to do these things at a great level. So, if we can just make it more accessible, suddenly goes from Codex is for coders to Codex is for everyone.

而在我们看到所有这些改进之后不久，硅谷出现了另一个现象，那就是 Open Claw。对吧，这可能是更广泛的科技社区，人们开始以你建议的方式信任它：让 AI 机器人访问他们的桌面，或者买一台 Mac Mini，让它访问邮件、日历、文件，然后让它管理他们的生活。然后 OpenAI 把 Open Claw 的创始人招入麾下。所以，你更多地谈到了 AI 作为一种帮助你管理生活的东西。把 Open Claw 团队招入麾下是出于这个愿景吗？

And soon after this moment where we saw all this improvement, there was another phenomenon in Silicon Valley, which was Open Claw. Right, which is maybe the broader tech community where people started to trust it in ways that you suggested: giving an AI bot access to their desktop, or getting a Mac Mini and giving it access to their mail, calendar, files, and then just kind of letting it go run their life. And then OpenAI brought the founder of Open Claw in-house. So, you talked a little bit more about the AI as something that will help run your life for you in a way. Is that the vision by bringing the Open Claw team in-house?

嗯，我认为这项技术的核心是弄清楚它如何有用，人们想如何使用它。智能体的愿景是什么？它将如何融入人们的生活？这是一个难题，我在多代技术中看到的是，那些真正投入、充满好奇心、有远见的人。这是一项真正的技能，也是正在兴起的新经济中非常有价值的技能。而 Open Claw 的创始人 Peter，我认为他拥有非凡的远见和创造力。所以，在某种程度上，这关乎具体技术，但在某种程度上又完全不是。它真正关乎的是我们如何利用这些能力，并弄清楚它们如何融入人们的生活。因此，作为技术专家，这非常令人兴奋，但作为专注于为人们带来实用性的人，这是我们正在加倍投入并大力投资的事情。

Well, I say that the core thing about this technology is that figuring out how it's useful, how people want to use it. What is the vision for agents? How is it going to slide into people's lives? That is a hard problem and one thing I've seen across many generations of this technology is the people who really lean in, who have a lot of curiosity, who have a lot of vision. That's a real skill and that's an emerging very valuable skill in this new economy that is emerging. And Peter, who is the Open Claw founder, is I think someone who's got incredible vision, incredible creativity. And so, to some extent it's about the specific technology, but to some extent it's not at all. It's really about how do we take these capabilities and figure out how they slide into people's lives. And so, I think as a technologist, it's very exciting, but as someone who is focused on bringing utility to people, that's something that we are doubling down on and investing quite a lot.

你最近对此有一个相当有趣的引用。谈到让这些自主 AI 智能体为你工作。你说，当你这样做时，你成为了一个由数十万个智能体组成的舰队的 CEO，它们正在完成你的目标、你的愿景，而你并不深究不同事情是如何解决的。在某种程度上，这种新的工作方式会让你感觉失去了对问题的脉搏。这好吗？

You had a pretty interesting quote about this recently. Talking about getting these autonomous AI agents to work on your behalf. You said, you become when you do it, you become this CEO of a fleet of hundreds of thousands of agents that are completing your objectives, your goals, your vision, and you're not in the weeds on exactly how different things are solved. And in some ways, this new way of work can make you feel like you're losing your pulse on the problem. Is that good?

我认为好坏参半。所以，我认为我们需要做的是承认这些工具能带来的优势，并减轻其弱点。因此，给人们杠杆和能动性，让他们如果有愿景、有想完成的事情，就能拥有一支智能体舰队去为你完成。但是，如果你思考世界是如何运作的，最终总有一个责任方。对吧？如果你试图建一个网站，你的智能体搞砸了，用户受到影响，那实际上不是智能体的错，而是你的错。所以，你需要关心。

I think there's a mixed bag. And so, I think that what we need to do is acknowledge the strengths of what these tools can deliver and mitigate the weaknesses. And so, giving people leverage, agency, making it so that if you have a vision, something you want to accomplish, that you can have a fleet of agents that will go do it for you. But, if you think about how the world works, at the end of the day, there's an accountable party. Right? If you're trying to build a website and your agent messes it up, and your user is affected, it's not really the agent's fault, it's your fault. And so, you need to care.

我认为，要正确使用这些工具，你必须认识到人类能动性、人类问责制是系统的核心部分。人类如何使用 AI，这是非常根本的。所以，重要的是，作为这些智能体的用户——我们在 OpenAI 内部也这样做——你不能推卸责任。你不能只是说：“啊，AI 会搞定一切的。”

And I think that for people to use these tools right, you need to realize that human agency, human accountability, that's a core part of the system. How the human uses the AI, that's something that is deeply fundamental. And so, I think the important thing is that as a user of these agents, and we do this within OpenAI, you cannot abdicate responsibility. You cannot just say, "Ah, the AI is just going to do stuff."

当然，但你说你感觉对问题本身失去了把握。这和问责层面是不同的。

Of course, but you said you feel like you're losing your pulse on the problem itself. That's different from the accountability layer.

嗯，对我来说，它们实际上是联系在一起的，因为关键在于，如果你是一位 CEO，离细节太远，对吧？如果你在经营这家公司，管理这个团队，却失去了对脉搏的把握，那不会带来好的结果。所以，我想表达的是，人类不必知道所有事情并不是一件好事。有些细节，因为你可以信任——比如你和总承包商团队一起盖房子，有很多细节你可能不需要担心，因为你可以信任他们会处理好。但归根结底，如果细节出了问题，你应该在意。你应该意识到。所以，我认为这是一个重要的细微差别：你不能盲目地说，“我不介意失去对脉搏的把握。”我们需要投入进去，说，“我需要保持把握，才能真正理解优势和劣势。”当你脱离一些细节、这些低层次的机械性事务时，你应该是因为已经建立了对系统的信任，相信它能做好工作。

Well, to me, they actually are linked together because the point is that if you're a CEO and you're too far from the details, right? If you're running this company, you're running this team, and you've lost your finger on the pulse, that is something that's not going to lead to great results. And so, the point that I was trying to make there is that it's not a desirable thing for humans to not have to know about what's going on. There are some details that, because you can trust—like if you are working with a team like a general contractor to build a house, there's a bunch of details there that you probably don't need to worry about because you can trust that they'll be taken care of. But at the end of the day, if there are details that are wrong, you should care about it. You should be aware. And so, this is, I think, an important nuance: you cannot just blindly say, "I'm okay with losing my finger on the pulse." We need to lean in and say, "I need to keep it there to really understand the strengths and weaknesses." And as you disengage from some of these details, these lower-level mechanical things, you should do it because you have built trust with a system that it will do a good job.

最后一个关于模型的问题。你谈到了一些模型经历的演变。预训练和微调、强化学习，使其更有能力逐步解决问题，并上网做事。现在我们正处于模型通过这个过程学会使用工具的时刻。如果我说错了请纠正我。那么，这个进展的下一步是什么？

One last question about the models. You talked a little bit about the evolutions that the models have gone through. Pre-training and fine-tuning, reinforcement learning that gets it more equipped to solve problems step-by-step and go out on the internet and do things. And now we're in this moment where the models have learned through that process to use tools. And correct me if I'm wrong on this one. What is next in that progression?

嗯，我认为我们正处于机器能力与深度不断提升的世界。其中一部分是关于工具使用，但现在我们还需要真正构建出色的工具。想想像计算机使用这样的东西——一个能实际使用桌面的 AI，那么它就能做你能做的任何事情。但我们还需要为机器构建一些东西：想想企业认证如何工作？审计追踪和可观测性如何工作？

Well, I think that the world we're in is one of increasing capability and depth of what the machine can do. And some of this is about tool use, but now we also need to actually build really great tools. You think about something like computer use—an AI that can actually use a desktop, then it is really able to do anything that you can do. But we also have to build a bit for the machine: think about how enterprise credentialing works? How do audit trails and observability work?

嗯。

Mhm.

有很多技术需要构建，才能跟上核心模型的能力。我认为总体方向包括像非常出色的语音界面这样的东西。所以，你可以自然地与电脑对话，就像这次对话一样自然，它能理解你。它能做你需要的事。它能给出好建议。它能呈现出来——我一直在做这件事，我有个问题。这里——你早上醒来，它会说：“这是你的每日报告，你的智能体昨晚取得了多少进展。”也许它为你经营业务，我认为这将是这项技术的巨大应用。创业的民主化绝对会到来。它会说：“这里有这些问题。有个客户不高兴，你知道，他们想和真人说话。你应该去和他们谈谈。”所有这些都会发生。然后，我认为提高人类能解决的挑战的雄心上限，也是这项技术的下一步。我们正在看到它的前沿。我非常兴奋看到的东西，几乎就像——你还记得 AlphaGo 的第 37 步吗？那一步是人类永远想不出来的。

There's a lot of technology to build to catch up with what the core model capability is. And I think the overall direction of travel includes things like a really great speech interface. So, you can just talk to your computer naturally, just as natural as this conversation, and it understands you. It does what you need. It has good advice. It's able to surface that—I've been working on this thing, I have a problem. Here's—you wake up in the morning and it says, "Here's your daily report of how much progress your agents made overnight." Maybe it's running a business for you, which I think is going to be a huge application of this technology. The democratization of entrepreneurship is absolutely coming. It'll say, "Here are these problems. There's this customer that's upset, you know, they want to talk to a real human. You should go talk to them." All of that's going to happen. And then, I think that the raising of the ceiling of ambition of challenges humanity can solve is also a next step for this technology. And we're seeing the leading edges of it. The thing that I am just very excited to see is almost, if you remember AlphaGo move 37, right? That move that no human ever would have come up with.

它很有创意。

It's creative.

有创意。它改变了人类对围棋的理解。这将在每一个领域发生。它将发生在科学、数学、物理、化学中。它将发生在材料科学、生物学、医疗保健、药物发现中。但它甚至可能发生在文学、诗歌以及许多其他领域，以我们现在无法想象的方式解锁人类的创造性理解和构思。

Creative. And it changed humanity's understanding of the game. That is going to happen in every single domain. It will happen in science, in math, in physics, in chemistry. It's going to happen in material science, going to happen in biology, it's going to happen in healthcare, drug discovery. But it may also even happen in literature, in poetry, in a bunch of other fields that will unlock human creative understanding and ideation in ways we can't imagine right now.

既然你说模型如此强大，为什么你认为这还没有发生？

Why do you think that hasn't happened yet, given how strong you say the models are?

嗯，我认为存在一个差距，即模型的能力和人们如何使用它们之间。所以，应用——嗯，是的，这几乎是我们对这些模型内部的理解。

Well, I think that there is an overhang where what the models are capable of and how people are using them. So, the application—well, yeah, it's almost our understanding of what is in these models.

好的。

Okay.

我认为这仍在显现中。所以，即使没有进一步进展，仍然会发生巨大的转变。由计算 AI 驱动的经济仍将到来。但我认为还有一点：我们非常擅长训练模型处理可衡量的任务。所以，我们从数学问题开始。编程问题，你有完美的验证器。而很多进展在于将这一点扩展到更开放的问题，扩大了可创造的空间。AI 本身可以在这方面提供很大帮助。如果 AI 聪明且理解事物，你给它一个任务完成得如何的评分标准。当然，对于像创意写作这样的东西，比如“这是一首好诗吗？”那是更难评分的事情。所以，我们教 AI、让它体验和尝试的能力较弱。但这一切都在改变，我们对此有很多见解。

That's something that I think is still emerging. So, I think that even with no further progress, there's still a massive shift that will happen. The economy being powered by computing AI is still going to happen. But I think there's also something where what we've gotten very good at is training models on tasks that could be measured. And so, what we started with was math problems. Programming problems where you have a perfect verifier. And a lot of what the progress has been in bringing this to more open-ended problems has been expanding the space of what can be created. And the AI itself can really help with that. If the AI is smart and understands things, you give it a rubric for how well a task goes. And of course, for things like creative writing, like "Is this a good poem?" That's a much harder thing to grade. And so, we've had less ability to teach the AI and for it to experience and try things out. But all of that is changing and something that we have a lot of sight for.

现在，有趣的是，彼得·蒂尔曾提到——我很确定他是这么说的——如果你是数学人，在这些模型取代你的工作方面，你可能比文字人面临更大的麻烦。而你当年是数学俱乐部的成员。你不担心吗？

Now, it's interesting reflecting on that Peter Thiel has mentioned, pretty sure this is what he said, that if you're a math person, you're probably in deeper trouble in terms of these models coming for what you do than if you're a words person. And you were a member of math club back in the day. Are you not concerned about that?

嗯，我认为看到我们失去的东西比看到我们得到的东西容易得多，对吧？因为我们深刻理解“我以前是这样做的。我以前参加数学竞赛。现在 AI 能参加数学竞赛了。”但这从来不是关于数学竞赛本身。对吧。那并不是驱动人类的东西。如果你想想我们现在的工作方式——有一个盒子，我们在盒子后面打字。100 年前我们不是这样做的。那不是自然的。那是我们都陷入的数字世界。那并不是做人的意义。做人是关于在这里、在当下、与他人连接。

Well, I think that it's much easier to see what we lose than what we gain, right? Because we have a deep understanding of "I used to do things this way. I used to do this math competition. Now the AI can do the math competition." But it was never really about the math competition. Right. That's not really the thing that drives humanity. And if you think about the way that we do work right now—there's a box and we type behind a box. We weren't doing that 100 years ago. That's not natural. That's this digital world that we all got kind of sucked into. That's not really what being human is about. Being human is about being here, being present, connecting with other humans.

我认为我们将看到的是，AI 将释放大量时间，用于增进人与人之间的联系，建立更多纽带。这是我非常期待的事情。

And I think that what we're going to see is that AI is going to free up so much time to increase human connection, to build more bonds across people. And that's something I'm extremely excited about.

好的。当我们转向这些更智能体式的用例时，有人讨论是否真的需要更大的训练运行。特别是如果你把模型做得足够好，那么你可以让它走向世界，并在非预训练领域获得很多提升，而预训练正是这些大数据中心所需要的。你在这里负责 Scaling。领导这个过程。你怎么看这个论点？

Okay. And then as we shift to these more agentic use cases, there's been discussion about whether the bigger training runs really need to happen. Especially if you get the model good enough, then you could sort of let it go out in the world and effectively get much of the uplift in areas that aren't the pre-training, which is what these big data centers are needed for. So, you work on scaling here. Lead that process. What do you think about that argument?

嗯，我认为它忽略了技术发展过程中非常重要的一点，因为模型生产流程的每一步都是相乘的。所以，你想改进所有步骤。我们看到，改进预训练会使所有其他步骤变得更容易，这是有道理的，因为模型能够学得更快。一个模型，因为它在尝试不同想法并从自身错误中学习时已经更有能力，这个过程就会更快。它需要犯的错误更少。所以，我认为一个大的转变是，从仅仅认为你是在单独训练这个大脑系统，让它越来越大，转变为也要尝试新事物。还要了解人们如何在现实世界中使用它，并将其反馈到训练中，但这并没有消除继续这项研究的价值和重要性。另一个我认为已经改变的是，我们过去只关注原始的预训练能力，而不太考虑推理能力，这在过去 24 个月里是一个重大变化，我们意识到需要平衡：你可以拥有一个基础模型，它具备所有这些优秀特性，但你真的需要它能够进行推理，因为你需要做强化学习，你需要将它服务于世界。这意味着你不一定非要尽可能大，因为你还要考虑所有下游使用，你真正想要的是最佳智能乘以成本，并一起优化这两者。

Well, I think it misses something very important for how the technology development goes because it is absolutely the case that every single step of the model production pipeline multiplies. And so, you want to improve all of them. And the thing that we see as we improve the pre-training, it makes all the other steps much easier and it makes sense because a model is able to learn faster. It's a model that, because it is already more capable to start when it's trying out different ideas and learning from its own mistakes, that process just is faster. It needs to make fewer mistakes. And so, I think that the big shift has been from thinking of it as just you're training this cerebral system on its own and you just make it bigger and bigger to it's also about trying things out. It's also about understanding how people are using it in the real world and connecting that back into your training, but it doesn't remove the value and the importance of continuing that research. And the thing that I think has also shifted is we used to really just focus on the raw pre-training capability, but not think as much about the inference ability and that's been a big change over the past 24 months to realize that it's a balance between you can have this model that has all those great properties in the base, but then you really need it to be able to be inferencable because you need to do reinforcement learning, you need to serve it to the world. And that means that you don't necessarily go as big as you possibly could because you also really think about there's going to be all this downstream use and you really want the thing that has the best intelligence times the cost and to optimize those two things together.

如果事情主要转向推理，你还需要 Nvidia GPU 吗？

Do you still need the Nvidia GPU if things move mostly to inference?

我们绝对需要。是的。

We absolutely do. Yes.

为什么？

Why?

嗯，原因有很多，但其中之一是，即使推理与训练的比例发生变化，除了将算力集中在一个问题上，你无法通过其他方式获得大规模训练。所以，我认为会发生的情况是，部署规模会大幅增加，但有时会有一个特定的巨大预训练运行，你确实想在那里集中大量算力。我也认为 NVIDIA 团队非常出色，做了非常了不起的工作。所以，是的，我们与他们密切合作。

Well, because there are multiple reasons, but one is that even as the balance of how much inference versus training changes, you cannot get massive scale training through any other way besides this concentration of compute on one problem. And so, I think that the thing that will happen is there's some amount of the deployment footprint goes up quite a lot, but that sometimes there will be a particular massive pre-training run, and you really want to concentrate a bunch in there. I also think that the NVIDIA team is just incredible and does really, really amazing work. And so, yeah, we partner very closely with them.

难道不会有那么一天，人们只是说，“我们预训练够了。模型足够聪明了？”

Isn't there going to be a time where people just say, "We've pre-trained enough. The models are smart enough?"

我认为这有点像，一旦人类解决了我们面前的所有问题，那么也许我们可以这么说。对吧。但我认为我们想要达到的上限，我认为有太多的雄心壮志，也许我们在过去 50 年左右的时间里有点退缩了，对吧？你想想，即使是看起来非常明确的问题，比如我们能否让每个人都拥有医疗保健，不仅仅是针对人们有问题的时候，而是真正考虑生活方式，以及如何真正帮助人们及早发现潜在疾病。我认为通过更智能的模型，我们实际上可以实现这个问题。也许在某个水平上，你可以完全解决这个问题，然后你会说，“嗯，我需要一个聪明两倍的模型吗？”但还有其他问题会需要它。

I think that's a little bit like once humanity has solved all problems in front of us, then maybe we can say that. Right. But I think that the ceiling of what we want to accomplish, I think that there's just so much ambition that maybe we've over the past 50 years or so just sort of backed off from, right? You think about even problems that seem very clear like can we have healthcare for everyone that is not just targeting when people have a problem, but really think about the lifestyle and how to really help people early detect potential diseases before they happen. Like that's a problem that I think we can actually achieve through more intelligent models. And there's probably some level where you can totally solve that problem, and then you say, "Well, do I need a model that's two times smarter?" But there are other problems that are going to demand that.

我们来谈谈建造这些数据中心的数学问题。你今年早些时候筹集了 1100 亿美元。这背后的数学是什么？这笔钱直接投入数据中心吗？你如何考虑如何向投资者返还这笔钱？谈谈这些计算。

Let's talk about the math about building these data centers. You raised 110 billion earlier this year. Is what's the math behind that? Does that money go right into data centers? How do you think about how you're going to return that money to investors? Talk about those calculations.

是的，所以我认为很简单，我们面前巨大的开支就是算力。但你可以把算力看作不是成本中心，而是收入中心。有点像雇佣销售人员。对吧？你想雇佣多少销售人员？只要你能销售你的产品，只要你有可扩展的方式来销售产品，那么销售人员越多，你赚的钱就越多。我认为我们所处的世界是，我们不断发现我们无法足够快地建造算力来满足需求。我非常具体地看到了这一点。对吧？现在我们不得不做出非常痛苦的决定，关于我们推出什么，算力流向哪里，我认为随着我们转向这个 AI 驱动的经济，我们将在经济中更广泛地经历这一点。问题将是什么问题会获得那些巨大的算力。你如何扩展，以便每个人都能拥有一个为他们运行的个人智能体？每个人如何能使用像 Codex 这样的系统？世界上根本没有足够的算力来做到这一点。所以，我们正试图领先于这个问题。但是，这是一个新的类别。

Yeah, so I think it's as simple as the massive expense we see in front of us is compute. But you can think of compute not as a cost center, but as a revenue center. Think of it a little bit like hiring sales people. Right? How many sales people do you want to hire? As long as you can sell your product, as long as you have a scalable way to sell that product, then the more sales people you have, the more revenue you will make. And I think the world that we're in is we have continually found we cannot build compute fast enough to keep up with demand. And I see this very concretely. Right? Right now we have to make very painful decisions about what we're launching, about where the compute goes, and that I think we're going to experience this more broadly within the economy as we shift to this AI-powered economy. The question will be what problems are going to get that massive compute. How do you scale so everyone can have a personal agent running for them? How can everyone be using systems like Codex? Like there just isn't enough compute in the world to be able to do that. And so, we're trying to get ahead of that problem. But, it is a new category.

对吧？所以，你以真正的信心在做这件事。我的意思是，投入这个项目的资金规模是世界从未见过的。当你建立一个新类别时，你如何确信它会成功？

Right? So, you're doing it with real confidence. I mean, in sums of money the world has never seen put towards a project like this. When you're building a new category, how do you do it with certainty that it's going to work out?

嗯，我认为有几个因素。首先，目前有历史先例。从我们推出 ChatGPT 的那一刻起，我记得我和我的团队有过这样的对话。他们说，“我们应该买多少算力？”我说，“全部。”他们说，“不，不，不，说真的。我们应该买多少算力？”我说，“无论我们尝试建造多少，我知道我们无法跟上需求。”这是真的，从那以后每年都是如此。挑战在于，这些算力采购，你必须在机器交付前 18 个月，有时 24 个月，有时更长时间锁定，这意味着你真的需要向前预测。

Well, I think there are several components that go into it. So, the first is there is historical precedent at this point. From the moment we launched ChatGPT, I remember talking with my team having this exact conversation. They said, "How much compute should we buy?" I said, "All of it." They said, "No, no, no, really. How much compute should we buy?" I said, "No matter how much we try to build, I know we're not going to be able to keep up with the demand." And that has been true, and that has been true every year since then. And the challenge is that these compute purchases, you have to lock them in 18 months, sometimes 24 months, sometimes longer in advance of the machines being delivered, which means you really need to project forward.

我认为我们正在走向的世界是，迄今为止我们的大部分收入来自消费者订阅，这一点将永远非常重要。我们还有其他收入来源正在涌现，但现在明显出现的机会是知识工作。我们非常具体地看到，每一家企业都意识到这项技术确实有效，为了保持竞争力，他们需要采用它。你可以看到这种自发的能量，所有软件工程师都在使用它，然后我们开始看到人们在企业内部将其用于各种知识工作，支付意愿和收入增长在这个行业非常明显。对吧？这正在发生，只需向前看。我们能看到的一件事，也许世界看不到的，是这些模型将如何改进的路线图。所有这些加在一起表明，经济是一个巨大的东西，对吧？经济如此之大，几乎无法理解。所有的增长，从最高层面看，经济从这里增长的方式将取决于 AI，取决于你如何利用 AI 以及你拥有的算力来驱动它。

And I think that the world that we're moving towards is one where to date most of our revenue has come from consumer subscriptions, and that will always be very important. There's other revenue streams we have emerging as well, but the opportunity that clearly is emerging now is knowledge work. And we're seeing this very concretely across every single enterprise realizing this technology it actually really works and to be competitive they need to adopt it. And you can see this organic energy of all these software engineers using it and then we're starting to see the percolation of people using it for various knowledge work inside of the enterprises and the willingness to pay and the revenue growth that you're seeing in this industry is very clear. Right? It's very clearly happening right now and just project that forward. And we look like one thing we get to see that maybe the world doesn't is the line of sight to how these models will improve. And all of this together says that the economy, which is a massive thing, right? The economy is just so large it's almost incomprehensible. All of the growth, like the highest order bit on how this economy grows from here will be about AI, how well you can leverage AI and the computational power you have available to power it.

你说消费者订阅目前是你最大的收入来源。预计这种情况会翻转，企业业务会成为最大来源吗？

You said consumer subscriptions are your biggest source of revenue right now. Is the projection that will flip and that business will be the biggest source?

我认为，很明显企业——不仅仅是企业，因为企业也在改变其含义。所以实际上是人们将其用于生产性知识工作这类事情。我认为在考虑定价时，如果你看看 Codex 目前的工作方式，如果你有 ChatGPT 消费者订阅，你就可以使用 Codex。所以我认为不会像有这一类、那一类那样定义明确。我认为这实际上关乎你作为用户，就像你的笔记本电脑一样，拥有这个通往数字世界的门户，而收入从根本上将来自这里。

I think well, I think that it is very clear how quickly the enterprise, it's not just enterprise because I think enterprise is also changing what it means. So really people using it for productive knowledge work for those kinds of things. And I think that as we think about pricing, one thing if you look at how Codex works right now is if you have a ChatGPT consumer subscription, you can use Codex. And so I think it's not going to be as well defined as there's this category, that category. I think it will really be about you as a user are going to have just again like your laptop this portal to the digital world and that is what the revenue fundamentally will come from.

Dario 说，‘我想到了你。’有些玩家在冒险，把旋钮拧得太过了，我非常担心。我认为他指的是你们的基础设施赌注。你怎么看？

Dario said, 'I think about you.' There are some players who are yoloing who pull the wrist dial too far and I'm very concerned. I think he's referencing your infrastructure bets there. What do you think about that?

嗯，我不同意。我认为我们一直非常深思熟虑，并且非常清楚即将发生的事情，我认为我们甚至会在今年看到每个参与者都会面临算力紧张。我认为我们是最先认识到这一点并提前建设以应对技术发展的。我认为我们看到的是，其他玩家可能去年年底才意识到这一点，然后开始争抢可用的算力，但真的没有了。所以我认为，即使人们很容易发表这样的言论，但每个人都意识到这项技术是有效的，它就在这里，是真实的，对吧？软件工程只是第一个例子。我们从根本上受限于可用的算力。

Well, I just disagree. I think we've been very thoughtful and very much seeing what is coming and I think that we will see even this year how everyone who is participating is going to be compute strapped. And I think we have been the most forward in realizing that this is coming and building in anticipation of how this technology is playing out. And I think that what we have seen is that for other players that they kind of realized that probably late last year and started scrambling to see what compute is available and there really wasn't any. And so I think that even as people it's very easy to make statements like that, but I think that everyone has kind of realized that this technology it's working, it's here, it's real, right? Software engineering is just the first example of it. And that we are fundamentally limited by the computational power available.

你还说过，如果他的预测稍有偏差，他的公司可能会破产。你也是这样吗？

And you said that also that if he's off by with his prediction by a little bit then his company could potentially go bankrupt. Is that the same case for you?

我认为，实际上这里有更多的退出路径。

I think that Look, I think that there's actually more degrees of off-ramp here.

好的。

Okay.

如果你开始担心下行风险，我认为这是一个非常合理的问题，对吧？但在某种程度上，我认为赌注不是关于任何一家公司。它实际上是关于整个行业。它实际上是关于你是否相信这项技术能够被生产出来，并交付我们看到即将到来的巨大价值。再次，我会指出证据点，对吧？软件工程就像，如果你不是软件工程师，你没有尝试过代码应用，那么它的不同之处简直难以描述。我认为人们会很快体验到。比如，6 个月前，我认为我们在内部看到了这一点，但外部证据点较少。现在有证据点了。6 个月后，我认为每个人都会感受到，我们都会感受到痛苦：有一个很棒的模型，但没有可用性，因为没有足够的算力。

If you start to worry about the downside case, which I think is a very reasonable question, right? But to some extent what I think the bet is on isn't about anyone company. It's really about the sector. It's really about do you believe this technology can be produced and can deliver this massive amount of value that we see coming. And again, I'll point to proof points, right? That software engineering it's just like the degree to which if you're not a software engineer, you haven't tried code apps the degree to which it's different, like it's just hard to describe. And I think that people will experience it very quickly. Like, you know, 6 months ago I think that for us we saw this internally, but there were less proof points out there. Now there's proof points out there. 6 months from now, I think that everyone will feel it and I think that we will all feel the pain of there's an awesome model and there's just no availability because there's not enough compute.

是的，但当我们在这期节目中展望 2026 年的预测时，去年年底我们有一次对话，Runa John Roy 和我们一起说，2026 年将是每个人都使用智能体的一年，我说，‘嗯，好吧，眼见为实。’而现在我正在使用智能体。所以。

Yeah, but as we were looking at our predictions for 2026 on this show, we had a conversation towards the end of the year last year where Runa John Roy was on with us was like 2026 is going to be the year where everybody uses agents and I said, 'Yeah, well, I'll believe that when I see it.' And I'm using the agents. So.

我们到了。开始吧。你用它做什么？

Here we are. Here we go. What do you use it for?

我用它来为我的同事构建内部工具，让大家在视频发布时间和缩略图外观上保持一致，我还整合了 YouTube 的数据，这样我们就可以根据缩略图对视频表现进行排名，就像是一个我永远不会花钱购买的定制软件。我认为这个时刻有趣的一点是，软件可以规模化，被大众使用，但当你使用时，会有很多不适合你的东西。也许这让我们能够以更自然的方式与软件互动。

I use it to build internal tools for the people I work with to get on the same page about when videos are coming and what the thumbnails need to look like, and I'm also integrating things from YouTube so we can basically rank how the videos are doing based on thumbnail and like a custom-built piece of software that I never would have paid for. And that's one of the things that I think is interesting about this moment, I guess, is that software it scales, used by the masses, but when you use it, therefore, there's going to be so many things that are not made for you. And maybe what this does is it allows us to interact with software in a way that's much more natural.

我认为这是关键。再次，我经常思考这样一个事实：我们构建计算机的方式真的把我们拉进了这个数字世界。想想你花了多少时间在手机上滑动。

I think that is the key. And again, I just think a lot about the fact that the way we've built computers has really pulled us into this digital world. You think about how much time you just spend scrolling through your phone.

是的。

Yep.

对吧？你花多少时间点击不同的按钮，试图把这个东西连接到那个东西。比如，为什么？你为什么要这样做？相反，AI 应该让机器更接近你，为你个性化，理解你想要完成什么。我们有所有这些流行文化，你可以和电脑说话，它们会为你做事，而这开始成为现实。它开始成为你真正可以做的事情。我认为它的神奇之处在于，你必须亲自尝试才能真正理解。所以，我绝对认为我们正处于一个非常特殊的时刻。是的。

Right? The amount of time that you spend clicking different buttons and trying to connect this thing to that thing. Like, why? Why do you have to do that? Instead the AI being about bringing the machine closer to you, personalizing to you, understanding what you're trying to accomplish. And that we have all of this pop culture of just computers you can talk to and that they go and do stuff for you and it's starting to become real. It's starting to become the thing that you can actually do. And I think that the amazingness of that is something where you just have to try it to really understand. So, I definitely think it's a very special moment we're in. Yeah.

那么我想知道为什么 AI 在公众中如此不受欢迎？例如，YouGov 说，认为 AI 对社会影响负面的人数是对正面影响预期的三倍。

Then I want to know why is AI so unpopular with the public? YouGov, for instance, says three times as many Americans expect the effects of AI on society to be negative as they expect it to be positive.

我的意思是，你认为这背后的原因是什么？你担心 AI 的品牌形象吗？

I mean, what do you think the reasoning is behind that? And are you concerned about AI's brand?

嗯，我认为我们需要向国家展示 AI 对人们的好处。不仅仅是宏观经济增长或 GDP 提升，而是它如何改善人们的生活。我每天都能听到很多具体的故事。比如，有一个家庭，孩子头痛、有医疗问题，被拒绝做核磁共振，他们用 ChatGPT 研究症状，意识到可以向保险公司争取做核磁共振。他们这么做了，结果发现孩子有脑瘤，他们因此救了他的命，因为用 ChatGPT 获得了正确的信息。这只是一个故事。还有更多类似的故事，人们的生活通过使用这项技术、真正与技术合作而得到深刻改善或拯救。我认为这些故事没有被传播出去。这发生在很多人的生活中，但不知何故，这些故事还没有被讲述。我注意到，从 90 年代开始，很多流行文化对 AI 持非常负面的态度，担心可能出错。但当人们真正使用 AI 时，他们发现它的实用性和价值。所以，我非常担心我们未能成功帮助人们理解为什么这波技术浪潮会改善他们的生活。它将有助于改善人际联系，这是我非常关注的一点。如果你考虑这里的机会以及 AI 为何如此重要，我认为这将是未来经济和国家安全的源泉。这将关乎国家竞争力，而像中国这样的其他国家，AI 正朝着完全相反的方向发展。

Well, I think that there is something that we need to show the country of why AI is good for them. Not just for the broad economy, for growing the GDP and things like that, but how does it help them in their lives? And I think there are actually many very concrete stories that I hear every day. For example, there's a family where their child was having some headaches, some medical issues, was denied an MRI, and they researched the symptoms with ChatGPT and realized that they could make an argument to insurance to actually get the MRI. They did that, turns out he had a brain tumor, they were able to save his life because they used ChatGPT to get access to the right information. That's just one story. There are so many more just like that of people who have been deeply, profoundly, their lives have been improved or saved through their use of this technology, through partnering with the technology in a real way. And so, that is a story I don't think gets out there. I think that this is happening in so many people's lives, but somehow the story is not yet told. And one thing I notice is that there's a lot of pop culture from the '90s, from the historical context that we have, that's very negative on AI, that worries about what could go wrong. But when people actually use AI, they find utility in it, they find value in it. And so, I think that I am definitely very concerned about us not having successfully helped people understand why this technology wave is something that will improve their lives. It will help improve human connection, and that is something that's a big focus in my mind. And if you think about the opportunity here and why AI is so important, I think this will be the source of economic and national security going forward. I think it's going to be about national competitiveness, and that there are other countries like China where AI pulls in the exact opposite direction.

所以我在那里吃了。

And so I ate there.

是的，我认为我们承认这一点并真正理解如何让每个人受益非常重要。

Yes, I think it's very, very important that we acknowledge that and we really understand how to get the benefits for everyone.

但我们也处于一个政治不稳定的时期。人们对工作有担忧。每次我和别人谈论 AI，他们都会问：‘我的工作还能干多久？’然后当我想到数据中心时，民意调查甚至比 AI 整体更糟。这是皮尤的数据。更多人认为数据中心对环境、家庭能源成本和附近居民生活质量弊大于利。所以，我们现在处于好工作难找的时期，人们看到这些数据中心进入他们的社区，他们说这对环境、家庭能源和生活质量不好。我的意思是，他们错了吗？

But we also are in a time that's like politically unstable. There's concerns about work. People are, every time I speak with someone about AI, they're like, 'How long do I have left to work in my job?' And then when I think about the data centers, I mean, the polling is even worse than AI in general. This is from Pew. Far more people say data centers are mostly bad than good for the environment, home energy costs, and quality of life of those nearby. So, we are at this moment where good jobs are tough to come by, and people see these data centers come into their communities, and they say not good for the environment, home energy, and quality of life. I mean, are they wrong?

嗯，我认为关于数据中心肯定有很多错误信息。一个很好的例子是用水量。如果你看看我们的阿比林设施，它是世界上最大的超级计算机之一，它的用水量相当于一个家庭一年的用水量。对吧？所以用水量真的微不足道。然而有很多错误信息说这些数据中心消耗大量水。同样，在电力方面，我们承诺自己承担成本，不会推高人们的能源价格。现在整个行业都在做出这些承诺，因为改善当地社区非常重要。当我们建设数据中心时，我们真正努力进入当地社区，了解当地情况，以及我们如何提供帮助。这些数据中心带来税收收入，我认为它们创造了就业机会。它们带来了很多好处。所以我认为这关乎我们如何表现，我们非常认真地对待这一责任。

Well, I think there's definitely a lot of misinformation about data centers. Good example is water usage. If you actually look at our Abilene facility, which is one of the if not the biggest supercomputers in the world, the amount of water it uses is the same as a household over the course of a year. Right? So, it's really negligible water use. And yet there's a lot of misinformation that these data centers consume a lot. And so, similar on power, we have a commitment that we are going to pay our own way to not drive up energy prices for people. That's something that as an industry now that people are making these commitments because it is very important that we improve local communities, and when we build data centers, we really try to go into those local communities, understand what's happening on the ground, how we can help. There's tax revenue that are associated with these data centers, and I think that there's jobs that they create. There's a lot of benefits that come from them. And so I think that's one thing where it is about how we show up, and that's a responsibility that we take very seriously.

好吧，但如果他们的电力成本不会上涨，你就得引入电力，这可能意味着更多的污染。这难道不是问题吗？

Okay, but also like if their power costs are not going to go up, you have to bring in power, which means potentially more pollution. Is that not a concern?

嗯，我认为在不让能源成本上涨方面有更多细微差别。如果你看看今天的电网如何运作，实际上有很多闲置电力，这些电力存在但没有被利用，你需要升级输电系统，同样，把这部分成本放在我们身上而不是放在费率支付者身上非常重要。没错。很多地方有清洁电力，但实际上未被充分利用，几乎被浪费了。所以，有真正的理由升级电网（许多地方已经老化过时）会带来很多好处。这实际上对社区有真正的好处。比如我们在北达科他州看到，人们的费率下降了，因为数据中心出现并帮助改善了所有人的公用事业。

Well, I think that there's much more nuance in terms of not driving up energy costs. If you look at how the grid works today, there's actually a lot of stranded power, power that is there that is not being utilized, and that you need to upgrade the transmission systems, and again, that's something where putting that on us rather than putting it on the rate payers is very important. Right. There's lots of places where they have clean power that is actually being underutilized and just being kind of thrown away. And so, there's a lot of benefit that comes from having real reasons for that grid, which is aging and obsolete in many places, to upgrade. And that's something actually has real benefits to the community. Like we've seen, for example, in North Dakota that people's rates have gone down because a data center has shown up and has helped with improving the utilities for everyone.

好了，最后一个关于政治的问题。你向 Magna Inc.捐赠了 2500 万美元，这是一个支持特朗普的政治行动委员会。你与《连线》杂志谈过这件事，你说：‘任何我能做的支持这项技术惠及所有人的事情，我都会去做。’你知道，如果这让你成为一个单一议题选民或单一议题政治支持者，我想分享一个我一直好奇的问题：归根结底，一个更强大的国家不是会让你的目标更可行吗？即使候选人并不完全支持你所做的事情。难道一个更强大的国家，无论如何，不应该是任何政治活动的北极星吗？如果是这样，那么这是捐赠的一部分吗？

All right, one last question on the politics. You gave $25 million to Magna Inc., which is a PAC that is pro-Trump. You spoke with Wired about it, and you said, 'Anything I can do to support this technology benefiting everyone is a thing I will do.' And, you know, if that makes you a one-issue voter or one-issue political supporter, I'll share the one thing I always wonder when it comes to just this one-issue camp is ultimately, doesn't a stronger country make you, you know, sort of make your goals much more feasible. Even if a candidate isn't fully in support of what you're doing. Like, shouldn't a stronger country, no matter what, be the North Star of any political activity? And if that's the case, then is that part of the donation?

所以，我对此的看法是，我和我的妻子做了这笔捐赠。我们也向两党超级政治行动委员会捐款。我认为这项技术发展迅速。在未来几年内，它将真正改变一切。它将成为经济的基础，但目前并不受欢迎。我们真的想支持那些真正拥抱这项技术、真正参与其中的政治家。所以，我认为这项技术当然关乎提升我们国家。你知道，我是一个单一议题捐赠者。我觉得在这方面我可以做出独特的贡献。但这实际上只是表达对这项技术的支持，我们作为一个国家应该拥抱它。

So, the way I look at this, my wife and I made that donation. We've donated to bipartisan super PACs as well. I think this technology is one where it's coming quickly. Within the next couple years, it's really going to transform everything. Going to be the underpinning of the economy, and it's not popular. And we really want to support politicians who really lean into this technology, really engage with it. And so, I think that certainly this technology is about uplifting us as a country. I, you know, I am a one-issue donor. This is something where I feel like I have a unique contribution to make. But, it's really about just expressing support for this technology is something that we should be leaning into as a country.

你会对害怕 AI 的人说什么？我的意思是，如果你有机会直接对他们说。他们可能认为 AI 会抢走他们的工作。

What would you tell someone who's scared of AI? I mean, if you have a moment here where you can speak directly to them. They might think it's going to take my job.

它会污染我的社区。它会改变世界太快了。你对此有什么想说的？

It's going to pollute my community. It will change the world too fast. What's your message to that?

第一件事是尝试这些工具。因为，真正理解它能为你做什么，只有通过体验当前存在的 AI 才能真正体会到。我们今天看到这项技术带来了如此多的机会、潜力和赋能。你刚才谈到了一些现在可以构建的东西。

Number one thing is try the tools. Because, to really understand what it can do for you is something that only by experiencing the AI as it exists now will that really hit home. And we see so much opportunity and potential and empowerment coming from this technology today. You talked a little bit about what you can build now.

对。

Right.

以前从未建过网站的人现在可以建网站了。如果你想创办一个小企业，并且正在考虑所有后端处理以及如何实际管理它，所有这些事情，AI 现在就可以帮你。所以，我认为在你的生活中，思考它如何能帮助你的健康，如何能帮助你爱的人，如何能帮你赚钱，如何能帮你省钱，这些都将摆在桌面上。我认为看到将要改变的东西比看到你将获得的东西要容易得多。但我认为值得给它一个公平的机会，真正去理解方程的两边。顺便说一句，这是民调数据中没有被提及的一点。那些看到别人使用但自己没有尝试过的人，或者从未尝试过 AI 的人，要负面得多。而重度用户，甚至偶尔使用的人，通常对这项技术相当积极。就我自己而言，我们已经思考这项技术很长时间了。我看到在我们面前展开的景象比我们想象的更惊人、更有益，并且将产生比我们想象中更积极的影响。

People who have never built a website before can build a website. If you want to start a small business, and you're thinking about all the back-end processing and how to actually manage it, all those things, the AI can help you with that right now. And so, I think that in your life, thinking about how it can help you with your health, how it can help your loved ones, how it can help you make money, how it can help you save money, these are all going to be on the table. And I think it is much easier to see what's going to change than it is to see what you're going to gain. But I think that it's worth giving it a fair shot of really trying to understand both sides of the equation. That's the one thing that doesn't get talked about in the polling data, by the way. It's the people that have seen it used but haven't tried it themselves or the people that have never tried AI are much more negative. And then you get to the power users and are even people that use it casually and they're generally pretty positive about the technology. For myself, we've been thinking about this technology for a long time. What I see playing out in front of us is more amazing, more beneficial, and going to really have a much more positive impact than we ever imagined.

最后一个问题，你会如何建议某人为未来做准备？而且不能只是获取工具。我的意思是，我有朋友来找我，他们说，我不知道我的工作或世界会发生什么，我只是需要知道该怎么做。

So last one for you, how would you advise someone to prepare themselves for the future? And it has to be more than just getting the tools. I mean, I have friends who come to me and they say, I don't know what's going to happen with my job or the world and I just need to know what to do with this.

我确实认为第一件事是理解这项技术。我们看到的一件事是，那些从技术中获得最多的人，你必须带着好奇心去接近它，真正在你的工作流程中尝试它，真正能够克服最初的那个障碍：你有一个空白框，我该拿这个空白框做什么？对吧？真正培养这种能动性，这种“我可以是管理者”的感觉。我可以设定方向。我可以委派，我可以提供监督，真正培养这项技能，因为这将是最基本的。我们正在为人类构建这项技术，以帮助人类培养更多的人际联系，让人类能够花更多时间做他们想做的事。所以问题是，你想要什么？真正去明确这一点，并尝试借助这项技术实现它，将是最重要的事情。

I do think that the number one thing is about understanding the technology. One thing we've seen is the people who get the most out of the technology and you have to approach it with a curiosity trying to really try it in your workflows, really be able to get over that initial hump of you have a blank box, what do I do with a blank box? Right? To really develop this sense of agency, this sense of I can be the manager. I can set the direction. I can delegate, I can provide oversight and to really develop that skill because that is something that's going to be fundamental. Is we're building this technology for humans, to help humans foster more human connection, for humans to be able to spend more time doing what they want. And so the question is, well, what do you want? And really trying to crystallize that and trying to realize that with help of this technology is going to be the most important thing.

Greg，非常感谢你来做客。

Greg, thanks so much for coming on the show.

谢谢你邀请我。

Thank you for having me.

好的。感谢大家的收听和观看，我们下次在 Big Technology Podcast 再见。

All right. Thank you everybody for listening and watching and we'll see you next time on Big Technology Podcast.